Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
Abstract II
Molecular genetic approaches to decrease mis-
incorporation of non-canonical branched chain
amino acids into a recombinant protein in
Escherichia coli
Ángel Córcoles García - Dissertation
Molecular genetic approaches to decrease mis-incorporation of
non-canonical branched chain amino acids into a recombinant
protein in Escherichia coli
vorgelegt von
M. Sc.
Ángel Córcoles García
ORCID: 0000-0001-9300-5780
von der Fakultät III-Prozesswissenschaften
der Technischen Universität Berlin
zur Erlangung des akademischen Grades
Doktor der Naturwissenschaften
- Dr. rer. nat. -
genehmigte Dissertation
Promotionsausschuss:
Vorsitzender: Prof. Dr. Juri Rappsilber, Institut für Biotechnologie, TU Berlin, Berlin
Gutachter: Prof. Dr. Peter Neubauer, Institut für Biotechnologie, TU Berlin, Berlin
Gutachter: Prof. Dr. Pau Ferrer, Universitat Autònoma de Barcelona, Bellaterra
(Cerdanyola del Vallès), Spain
Gutachter: Dr. Heinrich Decker, Sanofi-Aventis Deutschland GmbH, Frankfurt am Main
Tag der wissenschaftlichen Aussprache: 11. Dezember 2019
Berlin 2020
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
I Abstract
Abstract
The incorporation of non-canonical branched chain amino acids (ncBCAA) such as norleucine,
norvaline and β-methylnorleucine into recombinant proteins during E.coli production processes has
become a crucial matter of contention in the pharmaceutical industry, since such mis-incorporation
can lead to production of altered proteins, having non optimal characteristics. Hence, a need exists
for novel strategies valuable for preventing the mis-incorporation of ncBCAA into recombinant
proteins.
This work presents the development of novel E. coli K-12 BW25113 strain mutants allowing
exogenous tunable expression of target genes involved in the BCAA biosynthetic pathway. For that
purpose, following single genes were knocked out by homologous recombination in order to
eliminate endogenous genetic expression: thrA, ilvA, leuA, ilvIH, ilvBN, ilvGM and ilvC. Expression
regulation of previously knocked-out genes was carried out by transforming arabinose-based tunable
expression plasmids containing the native sequence of target genes into the generated E. coli KO
mutants. The engineered E. coli hosts were screened in a mini-reactor system in fed-batch mode
under both standard cultivation conditions and under conditions reproducing large-scale effects.
During screening three different concentrations of L-arabinose were tested for each strain in order to
trigger different expression levels of a target gene. Screening was performed by comparing the
impurity profile of the recombinant mini-proinsulin expressed in each tested strain with the non-
engineered E. coli host. After screening, potential E. coli mutants showing the most significant
ncBCAA reduction were scaled-up to a 15L stirred tank bioreactor and analysed to confirm their
behaviour.
Screening experiments carried out in the mini-reactor system indicated that an up-regulation of ilvC,
ilvIH and ilvGM and down-regulation of leuA and ilvBN trigger a reduction of norvaline and
norleucine biosynthesis and mis-incorporation into mini-proinsulin. Results for ilvA and thrA suggest
that the threonine pathway is not the one redirecting more metabolic flux to α-ketobutyrate.
Moreover, it is noteworthy that norleucine was the most mis-incorporated ncBCAA and β-
methylnorleucine levels did not significantly change under tested experimental conditions.
Furthermore, a novel cultivation strategy consisting of pyruvate pulsing and oxygen limitation was
demonstrated to successfully mimic large-scale effects since biosynthesis of ncBCAA and metabolites
of the overflow metabolism were reported under those conditions. Among the tested genes, up-
regulation of ilvIH and ilvGM showed the highest reduction of ncBCAA biosynthesis and mis-
incorporation. Reproducibility of potential ilvIH and ilvGM-tunable E. coli strains was confirmed in the
15L reactor.These novel E. coli strains with a reduced ncBCAA mis-incorporation may have a crucial
role in the pharmaceutical industry since product quality is pivotal for commercial approval of
recombinant proteins that are to be used as human therapeutics.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Zusammenfassung II
Zusammenfassung
Der Fehleinbau von nichtkanonischen verzweigtkettigen Aminosäuren (ncBCAA) wie Norleucin,
Norvalin und β-Methylnorleucin in rekombinante Proteine während E. coli -
Proteinproduktionsprozessen ist in der pharmazeutischen Industrie zu einem entscheidenden
Kriterium geworden, da ein solcher Fehleinbau zur Produktion veränderter rekombinanter Proteine
mit suboptimalen Eigenschaften führen kann. Daher besteht ein Bedarf an neuen Strategien, um den
Fehleinbau von ncBCAA in rekombinante Proteine zu verhindern.
Diese Arbeit beinhaltet die Entwicklung von neuen E. coli K-12 BW25113-Mutantstämmen, in denen
die Expression von Zielgenen, die am BCAA-Biosyntheseweg beteiligt sind, extern gesteuert werden
kann. Zu diesem Zweck wurden folgende Gene mittels homologer Rekombination inaktiviert: thrA,
ilvA, leuA, ilvIH, ilvBN, ilvGM und ilvC. Gleichzeitig wurden diese Mutanten durch Plasmide
komplementiert, die das chromosomal inaktivierte Gen unter Kontrolle eines regulierbaren
Arabinose-Promoters enthalten. Diese E. coli-Stämme wurden in einem Mini-Bioreaktorsystem im
Fed-batch-Modus sowohl unter Standardkultivierungsbedingungen als auch im Scale-down-Simulator
mit pulsbasierter Zufütterung untersucht. Während des Screenings wurden für jeden Stamm drei
verschiedene Konzentrationen von L-Arabinose getestet, um unterschiedliche Expressionsniveaus
eines Zielgens auszulösen. Für jeden Stamm wurde dann der Fehleinbau nichtkanonischer
Aminosäuren in rekombinant exprimiertes Mini-Proinsulins quantifiziert und mit dem nicht-
manipulierten E. coli-Kontrollstamm verglichen. Nach dem Screening wurden E. coli-Mutanten mit
einem signifikant verminderten Fehleinbau zur Validierung der Ergebnisse im Laborreaktor unter
Scale-down Bedingungen getestet.
Bei dem im Mini-Reaktorsystem durchgeführten Screening zeigte eine Erhöhung der Expression von
ilvC, ilvIH bzw. ilvGM sowie eine Verminderung der Expression von leuA bzw. ilvBN eine Reduktion
der Norvalin- und Norleucin-Biosynthese und einen geringeren Fehleinbau in das Mini-Proinsulin. Die
Änderung der Expression der Gene ilvA und thrA zeigten keinen signifikanten Einfluss. Dies bestätigt
die Hypothese, dass die Synthese der nichtkanonischen Aminosäuren nicht über den Threoninweg,
sondern durch die direkte Kettenverlängeurng von Pyruvat verursacht wird.
Außerdem, ist es bemerkenswert, dass Norleucin die am häufigsten falsch eingebaute
nichtkanonische Aminosäure war und sich die β-Methylnorleucin-Spiegel unter getesteten
Versuchsbedingungen nicht signifikant änderten. Darüber hinaus konnte gezeigt werden, dass eine
neuartige Kultivierungsstrategie, die auf Pyruvat-Pulsen bei gleichzeitiger Sauerstofflimitierung
beruht, Scale-down-Effekte erfolgreich imitieren kann. Von den getesteten Zielgenen zeigte eine
Expressionssteigerung von ilvIH bzw. ilvGM den signifikantesten Effekt bezüglich einer verminderten
Synthese der nichtkanonischen Aminosären und dem Fehleinbau. Diese neuartigen E. coli-Stämme,
die einen verminderten Fehleinbau nichtkanonischer Aminosäuren zur Folge haben, können eine
entscheidende Rolle in der pharmazeutischen Industrie spielen, da die Produktqualität für die
Zulassung rekombinanter Proteine als Humantherapeutika von entscheidender Bedeutung ist.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
III List of Contents
List of Contents
Abstract .................................................................................................................................................... I
Zusammenfassung ................................................................................................................................... II
List of Contents ....................................................................................................................................... III
Acknowledgements .............................................................................................................................. VIII
List of Abbreviations and Symbols .......................................................................................................... X
Abbreviations ........................................................................................................................................ X
Symbols .............................................................................................................................................. XIII
1. Introduction .................................................................................................................................. 1
2. Literature Review .......................................................................................................................... 3
2.1 E. coli as industrial biofactory ....................................................................................................... 3
2.1.1 Advantages of E. coli as industrial biofactory ........................................................................ 3
2.1.2 Origin and genetic background of strain E. coli K-12 BW25113 ............................................ 3
2.1.3 Inclusion body-based recombinant protein expression in E. coli.......................................... 4
2.1.4 Insulin expression in E. coli .................................................................................................... 5
2.1.5 Physiological stress during recombinant protein expression in E. coli ................................. 7
2.2 Introduction to central E. coli metabolism ................................................................................... 8
2.2.1 Catabolite repression and oxidative respiration ................................................................... 8
2.2.2 Mixed-acid fermentation ...................................................................................................... 9
2.2.3 Overflow metabolism .......................................................................................................... 10
2.2.4 BCAA biosynthetic pathway ................................................................................................ 12
2.2.5 Regulation of the BCAA biosynthetic pathway ................................................................... 15
2.3 Non-canonical branched chain amino acids (ncBCAA) ............................................................... 23
2.3.1 Introduction to ncBCAA ....................................................................................................... 23
2.3.2 ncBCAA biosynthetic pathway ............................................................................................ 25
2.3.3 Substrate promiscuity of enzymes involved in ncBCAA biosynthesis ................................. 27
2.3.4 Conditions triggering ncBCAA biosynthesis ........................................................................ 29
2.3.5 Mechanism of ncBCAA mis-incorporation into proteins during translation ....................... 31
2.3.6 Inconvenients of ncBCAA biosynthesis and mis-incorporation into recombinant proteins 33
2.3.7 Strategies to avoid ncBCAA mis-incorporation into recombinant proteins ........................ 34
2.4 Recombineering in E. coli ............................................................................................................ 38
2.4.1 Homologous recombination based on FRT/FLP system ...................................................... 39
2.5 Tunable promoter systems in E. coli metabolic engineering ...................................................... 41
2.5.1 Introductory considerations ................................................................................................ 41
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
List of Contents IV
2.5.2 araC-PBAD promoter system ................................................................................................. 42
2.5.3 rhaBAD promoter system .................................................................................................... 43
2.5.4 lac promoter system ........................................................................................................... 44
2.5.5 Pm/XylS promoter system ................................................................................................... 46
2.5.6 Summary ............................................................................................................................. 46
3. Research Hypotheses and Aim of the Project ............................................................................. 47
4. Materials & Methods .................................................................................................................. 51
4.1 Materials, reagents and equipment ............................................................................................ 51
4.2 Software ...................................................................................................................................... 51
4.3 Bacterial strains ........................................................................................................................... 51
4.4 Plasmids ...................................................................................................................................... 51
4.5 Media .......................................................................................................................................... 51
4.6 Primers ........................................................................................................................................ 52
4.7 Standard molecular biology methods ......................................................................................... 52
4.7.1 PCR....................................................................................................................................... 52
4.7.2 Plasmid extraction (mini and midiprep) .............................................................................. 54
4.7.3 Genomic DNA extraction ..................................................................................................... 55
4.7.4 Preparation of E. coli electro-competent cells .................................................................... 55
4.7.5 Electroporation of DNA into E. coli electro-competent cells .............................................. 56
4.7.6 Curation of temperature-sensitive plasmids ....................................................................... 56
4.7.7 Electrophoresis analytical gel .............................................................................................. 57
4.7.8 PCR spin purification ........................................................................................................... 57
4.7.9 Extraction from DNA gels .................................................................................................... 57
4.7.10 Standard cloning based on restriction digestion and ligation ............................................. 57
4.7.11 InFusion cloning ................................................................................................................... 58
4.7.12 Mutagenesis ........................................................................................................................ 59
4.7.13 Sequencing .......................................................................................................................... 59
4.7.14 Preparation of bacterial glycerol stocks .............................................................................. 60
4.7.15 SDS-PAGE ............................................................................................................................. 60
4.8 Amino acid analysis by GC-FID .................................................................................................... 61
4.8.1 Intracellular soluble protein fraction and inclusion body isolation from cell extracts ....... 61
4.8.2 Hydrolysis of intracellular soluble protein fraction and inclusion bodies ........................... 62
4.8.3 ncBCAA analysis with EZ:faastTM for free (physiological) amino acid analysis by GC-FID . 62
4.8.4 ncBCAA analysis by gas chromatography-flame ionization detection (GC-FID) .................. 63
4.8.5 Data analysis of data generated by GC-FID ......................................................................... 63
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
V List of Contents
4.9 Generation of pSW3 plasmid variants with different LacI expression levels: pSW3, pSW3_lacI+
and pSW3_lacIq ........................................................................................................................... 64
4.9.1 Sequencing of pSW3 ............................................................................................................ 64
4.9.2 Cloning lacIq into pSW3 to generate plasmid pSW3_lacIq................................................... 64
4.9.3 Mutation of pSW3_lacIq to generate pSW3_lacI+ ............................................................... 65
4.9.4 Transformation of pSW3 into E. coli K-12 BW25113 and molecular verification ............... 66
4.9.5 Transformation of pSW3_lacIq into E. coli K-12 BW25113 and molecular verification ...... 66
4.9.6 Transformation of pSW3_lacI+ into E. coli K-12 BW25113 and molecular verification ...... 66
4.10 Design of an arabinose-based tunable expression plasmid (pACG_araBAD) ............................. 67
4.10.1 Features of the DNA fragments involved in InFusion cloning ............................................. 67
4.10.2 Generation of PCR fragments for InFusion cloning ............................................................. 67
4.10.3 InFusion cloning to generate plasmid pACG_araBAD ......................................................... 68
4.10.4 Cloning of genes into plasmid pACG_araBAD ..................................................................... 69
4.11 Design of an m-toluine-based tunable expression plasmid (pACG_XylSPm) .............................. 71
4.11.1 Generation of plasmid pACG_XylSPm ................................................................................. 71
4.11.2 Cloning of genes into plasmid pACG_XylSPm ..................................................................... 72
4.12 Development of strains to regulate expression of the BCAA biosynthesic genes (geneX-tunable
E. coli) .......................................................................................................................................... 73
4.12.1 Generation of strains E. coli K-12 BW25113 ΔgeneX (geneX: leuA, thrA, ilvA and ilvC)...... 74
4.12.2 Generation of strains E. coli K-12 BW25113 ΔilvIH and E. coli K-12 BW25113 ΔilvBN ....... 76
4.12.3 Transformation of pSW3_lacI+ into E. coli K-12 BW25113 ΔgeneX ..................................... 78
4.12.4 Transformation of pACG_araBAD_geneX into E. coli K-12 BW25113 ΔgeneX pSW3_lacI+
(geneX-tunable E. coli) ......................................................................................................... 78
4.12.5 Transformation of pACG_XylSPm_geneX into E. coli K-12 BW25113 ΔgeneX pSW3_lacI+ . 79
5. Results ......................................................................................................................................... 80
5.1 Analysis of mini-proinsulin expression in E. coli K-12 W3110M and E. coli K-12 BW25113
containing plasmid pSW3 ............................................................................................................ 80
5.1.1 Cultivation conditions.......................................................................................................... 80
5.1.2 Mini-proinsulin analysis with SDS-PAGE ............................................................................. 81
5.2 Analysis of mini-proinsulin expression in E. coli K-12 BW25113 containing different variants of
plasmid pSW3 .............................................................................................................................. 82
5.2.1 Cultivation conditions.......................................................................................................... 82
5.2.2 Mini-proinsulin analysis with SDS-PAGE ............................................................................. 82
5.3 Evaluation of L-arabinose induction in E. coli BW25113 ΔgeneX expressing pSW3_lacI+ and
pACG_araBAD_geneX (geneX-tunable E. coli) ............................................................................ 84
5.3.1 Cultivation conditions.......................................................................................................... 84
5.3.2 Gene expression analysis .................................................................................................... 84
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
List of Contents VI
5.4 Evaluation of m-toluate induction in E. coli BW25113 ΔgeneX expressing pSW3_lacI+ and
pACG_XylSPm_geneX (geneX-tunable E. coli)............................................................................. 86
5.4.1 Cultivation conditions.......................................................................................................... 86
5.4.2 Gene expression analysis .................................................................................................... 86
5.5 Establishment of a GC-FID method allowing analysis of canonical and non-canonical amino
acids ............................................................................................................................................ 87
5.5.1 Elaboration of calibration curves and determination of retention times for ncBCAA
norvaline, norleucine and β-methylnorleucine ................................................................... 88
5.5.2 Elaboration of calibration curves and determination of retention times for canonical
amino acids .......................................................................................................................... 88
5.5.3 Evaluation of the effect of hydrolysis on amino acid analysis ............................................ 90
5.5.4 Validation of the GC-FID method by analyzing a pure protein ........................................... 91
5.6 Establishment of cultivation conditions based on pyruvate pulsing leading to an increase of
ncBCAA mis-incorporation into recombinant mini-proinsulin expressed in E. coli at shake flask
level ............................................................................................................................................. 92
5.6.1 Cultivation conditions.......................................................................................................... 93
5.6.2 Analysis of ncBCAA .............................................................................................................. 93
5.7 Establishment of cultivation conditions based on pyruvate pulsing and O2 limitation leading to
an increase of ncBCAA mis-incorporation into recombinant mini-proinsulin expressed in E. coli
in a 10 mL mini-reactor and in a 15L reactor .............................................................................. 95
5.7.1 Experiment in a 10 mL mini-reactor .................................................................................... 95
5.7.2 Experiment in a 15 L reactor ............................................................................................. 102
5.7.3 Comparison 10 mL mini-reactor and 15 L reactor ............................................................ 111
5.8 Screening of geneX-tunable E. coli strains a mini-reactor system ............................................ 113
5.8.1 Cultivation conditions........................................................................................................ 113
5.8.2 Mini-proinsulin analysis by SDS-PAGE ............................................................................... 116
5.8.3 ncBCAA analysis ................................................................................................................. 116
5.9 Screening of potential ilvGM- and ilvIH-tunable E. coli strains in a 15 L reactor under conditions
triggering ncBCAA formation .................................................................................................... 121
5.9.1 Cultivation mode ............................................................................................................... 121
5.9.2 Mini-proinsulin analysis by HPLC....................................................................................... 125
5.9.3 Analysis of ncBCAA ............................................................................................................ 126
5.9.4 Acetate and formate analysis ............................................................................................ 128
6. Discussion .................................................................................................................................. 130
6.1 Analysis of mini-proinsulin expression in E. coli K-12 BW25113 containing different variants of
plasmid pSW3 ............................................................................................................................ 130
6.2 Evaluation of L-arabinose induction in E. coli BW25113 ΔgeneX expressing pSW3_lacI+ and
pACG_araBAD_geneX (geneX-tunable E. coli) .......................................................................... 131
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
VII List of Contents
6.3 Evaluation of m-toluate induction in E. coli BW25113 ΔgeneX expressing pSW3_lacI+ and
pACG_XylSPm_geneX (geneX-tunable E. coli)........................................................................... 132
6.4 Establishment of a GC-FID method allowing analysis of canonical and non-canonical amino
acids .......................................................................................................................................... 133
6.5 Establishment of cultivation conditions based on pyruvate pulsing and O2 limitation leading to
an increase of ncBCAA mis-incorporation into recombinant mini-proinsulin in E. coli ............ 134
6.6 Screening of geneX-tunable E. coli strains ................................................................................ 137
7. Conclusions and Outlook .......................................................................................................... 141
8. References ................................................................................................................................. 143
9. Theses ....................................................................................................................................... 160
10. Appendix ................................................................................................................................... 161
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Acknowledgements VIII
Acknowledgements
In the following lines I would like to show my appreciation to all people who contributed to the
successful accomplishment of this work.
Firstly, I want to express my sincere gratitude to my supervisors Prof. Dr. Peter Neubauer and Dr.
Peter Hauptmann for giving me the opportunity to perform my PhD at Sanofi-Aventis Deutschland
GmbH in close collaboration with Technische Universität Berlin (TUB). Their supervision, advice and
encouragement marked a milestone in how I perceive science nowadays. In the same vein, I would
like to thank Dr. Heinrich Decker and Dr. Sebastian Rissom for counting on me to develop my thesis
at Sanofi-Aventis Deutschland GmbH.
Secondly, I would like to specially thank all the colleagues from Sanofi-Aventis Deutschland GmbH,
who also contributed in various ways to this thesis. I thank Dr. Berndt Janocha, Klaus Sparwald and
Michaela Michel for supporting me concerning the use of GC-FID to perform amino acid analysis. I
thank Dr. Holger Penders, Dr. Manuel Krewinkel, Claudia Erhard, Frank Ludwig, Alexander Leske, Jose
Gonzalez Rubio, Thomas Trick and Melanie Hänig for their supervision and advice concerning
fermentations at reactor scale. I thank Dr. Simon Stammen, Dr. Hans-Falk Rasser, Karima Chaouki,
Sylvia Zink, Nicole Schottstedt, Anika Ludwig, Martina Hildebrand and Birgit Antonijs for all their help
concerning molecular biology. I would also like to express my gratitude to Janine Neumann and
Christiane Straub for helping me with the nightmare of the business trips and all the bureaucracy. I
also thank Waltraud König for inviting me to the regular lunch meeting for Spanish speakers, which
brought me the opportunity to feel a bit more home in a foreign country. In addition, I thank to
Reiner Olliger and Thomas Kuhl the possibility to perform HPLC analysis at their laboratories. I would
like to thank Sarah Charaf for her help regarding generation of strains E. coli K-12 BW25113 ΔilvBN
and E. coli K-12 BW25113 ilvG+ as well as plasmid pACG_araBAD_ilvBN. It was a pleasure to supervise
the development of her master thesis. But I want to particularly show my gratitude to Dr. Peter
Hauptmann and the colleagues Klaus Schaffer, Andre Urban and Dietmar Heep. It was a pleasure to
share lab with them for 3 years in such a privileged working environment. Their encouragement,
advice and sense of humor were one of the main pillars contributing to my motivation and the
success of this work.
Thirdly, I would like to express my gratitude to all the colleagues from Technische Universität Berlin
(TUB), who also contributed in various ways to this work. I specially thank to Emmanuel Anane, who
was my main guide during my secondment at TUB. It was a pleasure to learn more about E. coli
metabolism, fermentation technology and scale-down models after our interesting discussions. I am
happy that our collaboration could translate in a scientific publication. In the same way, I would like
to express my gratitude to Dr. Christian Reitz for showing me the homologous recombination
technology used in this thesis to perform gene knock outs. My next thanks goes to Prof. Dr. Peter
Neubauer, Dr. Stefan Junne, Benjamin Haby, Sebastian Hans and Florian Gauche for their scientific
imputs, discussions and experimental support during my secondment at TUB.
Forthly, I would also like to show my appreciation to the colleagues from Denmark Technical
University (DTU) who guided me during my secondment there, especially to Prof. Dr. Krist V. Gernaey
and Dr. Christian Bach.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
IX Acknowledgements
I am grateful for financial support from the European Union´s Horizon 2020 research and innovation
programme under the Marie Sklodowska-Curie grant agreement No. 643056 (Biorapid). In this
regard, I would like to thank the European Union institutions for making it possible for young
researchers to live such a rewarding experience. In the same way, I also express my gratitude to all
ESRs taking part in the Biorapid Consortium and to the project coordinator, Prof. Dr. Jarka Glassey,
for the pleasant moments we spent together during the training weeks and conferences celebrated
throughout the project.
Finally, I would like to thank my family and friends in Spain and all over the world who where always
encouraging and supporting me despite the distance. I also thank to all my friends in Frankfurt, my
cats and, specially, to Mario Hiebel, for his patience and persistent support during the hard times of
the PhD life. Thank you all for always being there and trust on me. You were my main motivation to
success in this work and do not give up. I love you! ¡Muchas gracias, os quiero mucho!
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
List of Abbreviations and Symbols X
List of Abbreviations and Symbols
Abbreviations
etc. et cetera
i.e. id est
e.g. exempli gratia
(d)NTP (deoxy)nucleoside triphosphate
2D 2-dimensional
3D 3-dimensional
AA amino acid
aaRS aminoacyl-tRNA synthetase
ABA L-2-aminobutyric acid
ACKA acetate kinase
ACS acetyl-CoA synthetase
ADH alcohol dehydrogenase
ADP adenosine diphosphate
AHAS acetohydroxy acid synthase
AK bifunctional aspartokinase/homoserine dehydrogenase 1
Ala alanine
Amp ampicillin
ASAD aspartate-semialdehyde dehydrogenase
Asn asparagine
Asp aspartic acid (aspartate)
asRNA antisense RNA
ATP adenosine triphosphate
BCAA branched-chain amino acids
bST bovine somatotropine
CaM mammalian calmodulin
cAMP cyclic adenosine monophosphate
CAP catabolite activator protein
cBCAA canonical branched-chain amino acids
CDW cell dry weight
CGSC E. coli Genetic Stock Center
CISY citrate synthase
cm, chlor chloramphenicol
CNBr cyanogen bromide
CNCl cyanogen chloride
CoA (CoA-SH) coenzyme A
CRP cAMP receptor protein
DH dihydroxy-acid dehydratase
DMSO dimethyl sulfoxide
DNA deoxyribonucleic acid
DO dissolved oxygen
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
XI List of Abbreviations and Symbols
dsDNA double-stranded DNA
EDTA ethylenediaminetetraacetic acid
EF-Tu elongation factor Tu
EI enzyme I
EII enzyme II
ER endoplasmid reticulum
FAD flavin adenine dinucleotide in oxidized form
FADH2 flavin adenine dinucleotide in reduced form
FdhF cytoplasmic molybdenum- and selenium-dependent formate dehydrogenase H
FHL formate hydrogenlyase complex
FID flame ionization detection
FLP site-specific flipase recombinase
FRT FLP recombinase recognition sites
GAPDH glyceraldehyde-3-phosphate dehydrogenase
GC gas chromatography
GDP guanosine diphosphate
Gln glutamine
GLU glutamate
Glu glutamic acid (glutamate)
Gly glycine
Gpp pppGpp-5’-phosphohydrolase
GTP guanosine triphosphate
His histidine
hM-CSF human macrophage colony stimulating factor
HPr histidine-containing protein
HSAT homoserine acyltransferase
HSD bifunctional aspartokinase/homoserine dehydrogenase 1
HSK homoserine kinase
Hyc hydrogenase component
IB inclusion body
IHF integration host factor
IL-2 interleukin-2
Ile isoleucine
ileRS isoleucyl-tRNA synthetase
IPMD 3-isopropylmalate dehydrogenase
IPTG isopropyl β-D-1-thiogalactopyranoside
IR ketol-acid reductoisomerase (NADP(+))
ISOM 3-isopropylmalate dehydratase
Kan kanamycin
kanR kanamycin-resistant
kanS kanamycin-sensitive
KB ketobutyrate
KGLU α-ketoglutarate
KIV α-ketoisovalerate
KO knock-out
KV ketovalerate
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
List of Abbreviations and Symbols XII
L-ara L-arabinose
LB Luria-Bertani
LDH lactate dehydrogenase
Leu leucine
leuRS leucyl-tRNA synthetase
lim. limitation
Lrp leucine-responsive protein
Lys lysine
MCS multicloning site
Met methionine
metRS methionyl-tRNA synthetase
NAD(P)+ nicotinamide adenine dinucleotide (phosphate) in oxidized form
NAD(P)H nicotinamide adenine dinucleotide (phosphate) in reduced form
ncBCAA non-canonical branched-chain amino acids
Ndk nucleoside diphosphate kinase
NL norleucine
Nva norvaline
OD optical density
OD600nm optical density at 600 nm wavelength
PCR polymerase chain reaction
PDHC pyruvate dehydrogenase complex
PEP phosphoenolpyruvate
PFL pyruvate-formate lyase
PFR plug flow reactor
Phe phenylalanine
PI pro-insulin
PMS, IPMS α-isopropylmalate synthase
ppGpp guanosine 5’-diphosphate, 3’-diphosphate
PPI pre-pro-insulin
pppGpp guanosine 5’-triphosphate, 3’-diphosphate
pro proline
PTA phosphate acetyltransferase
PTS PEP-dependent carbohydrate phosphotransferase system
PYR-O2 pyruvate pulsing and oxygen limitation
dH20 deionized water
RE restriction enzyme
Rel. q. relative quantity
RNA ribonucleic acid
RNAP RNA polymerase
rRNA ribosomal ribonucleic acid
RSH protein pair RelA/SpoT
SAM S-adenosylmethionine
SD, STD standard
SDS-PAGE sodium dodecyl sulfate–polyacrylamide gel electrophoresis
Ser serine
SOC Super Optimal broth with Catabolite repression
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
XIII List of Abbreviations and Symbols
ssDNA single-stranded DNA
ß-MNL ß-methylnorleucine
STR stirred tank reactor
Ta annealing temperature
TD threonine deaminase, encoded by ilvA
Thr threonine
Tm melting temperature
TMG methyl-1-thio-β-d-galactopyranoside
TrB transaminase B
TrC transaminase C
tRNA transference ribonucleic acid
Trp tryptophane
TS threonine synthase
TUB Technische Universität Berlin
Tyr tyrosine
UAS upstream activating sequence
Val valine
WT wild type
Symbols
µset set-point of the specific cell growth rate (h-1)
F feed flow rate over time (L h-1)
qs specific substrate uptake rate (gS gX-1 h-1)
S concentration of glucose (g/L)
t time (h)
V volume (L)
X biomass concentration (g/L)
σ32 heat-shock sigma factor 32
σs sigma regulator
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Introduction 1
1. Introduction
Biosynthesis and mis-incorporation of ncBCAA such as norleucine, norvaline and β-methylnorleucine
into recombinant proteins has been reported in a number of E.coli production processes. NcBCAA
can be synthesized by E. coli metabolism as byproducts of the BCAA biosynthetic pathway as a result
of the low specificity of the leu and ilv-operon-encoded enzymes for their substrates. There are three
main causes triggering ncBCAA biosynthesis in E. coli: overflow metabolism driven by inefficient
mixing in large-scale reactors, de-regulation of enzymes encoded by the leu operon resulting from
leucine depletion and deficiency of AHAS II activity. Free ncBCAA can be mis-incorporated into
cellular proteins through tRNA misaminoacylation during protein translation due to their structural
similarity with the respective canonical amino acids. Such mis-incorporation might lead to the
production of altered recombinant proteins, having non optimal characteristics e.g. altered biological
activity, modulated sensitivity to proteolysis and immunogenicity. This represents an important
concern for the pharmaceutical industry since product quality is pivotal for commercial approval of
recombinant proteins that are to be used as human therapeutics in order to ensure patient safety.
Several valuable strategies for reducing mis-incorporation of ncBCAA into recombinant proteins were
already described in the literature: mutation of methionine codons of the gene encoding the
recombinant protein, co-expression of enzymes capable of degrading ncBCAA, supplementation of
exogenous canonical amino acids, overproduction of methionine by mutating genes involved in
methionine and threonine biosynthesis, knocking-out genes involved in the BCAA biosynthetic
pathway, optimization of fermentation conditions, supplementation of trace elements and use of
alternative E. coli expression strains less prone to non-canonical BCAA mis-incorporation. However,
most of the aforementioned strategies present numerous disadvantages. This thesis focuses on
another approach which has not yet been explored and that might contribute to close the scientific
gap: engineering of novel E. coli strains allowing tunable expression of target genes involved in the
BCAA biosynthetic pathway.
Together with the novel Pm/XylS promoter, the classic araBAD promoter was also utilized in this
thesis in order to regulate expression of target genes involved in the BCAA biosynthetic pathway. In
order to ensure functionality of the araBAD promoter, strain E. coli K-12 BW25113 was selected as
the model organism for this study, since it is deficient in arabinose catabolizing enzymes.
Furthermore, in order to properly evaluate the effect that expression regulation of a target gene has
on ncBCAA biosynthesis and subsequent mis-incorporation into cellular proteins, mini-proinsulin was
selected as the model recombinant protein in this study. Protein sequence of recombinant mini-
proinsulin contains canonical amino acids (3 methionine, 14 leucine and 5 isoleucine residues) that
can be potentially substituted by the non-canonical counterparts (norleucine, norvaline and β-
methylnorlelucine, respectively) upon mistranslation. An available plasmid (pSW3) expressing
recombinant mini-proinsulin under the control of a Ptac promoter was tested in the genetic
background of E. coli K-12 BW25113, revealing leaky and weak expression. Thus, a number of pSW3
plasmid variants encoding for different levels of LacI repressor were tested in E. coli K-12 BW25113,
being plasmid pSW3_lacI+ the variant reporting the most optimal results. Furthermore, a method
allowing amino acid analysis (including ncBCAA) by gas chromatography (GC-FID) was successfully
established. In order to regulate expression of target genes involved in the BCAA biosynthetic
pathway it was first necessary to eliminate endogenous expression of such genes in the E. coli cell.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
2
Hence, following single target genes/operons were knocked out from the E. coli K-12 BW25113
genome by homologous recombination: thrA, ilvA, leuA, ilvIH, ilvBN, ilvGM and ilvC. Expression
regulation of previously knocked-out genes was carried out by transforming arabinose and m-
toluate-based tunable expression plasmids (pACG_araBAD and pACG_XylSPm series) containing the
native sequence of the target genes into the respective engineered E. coli K-12 BW25113 KO
mutants. Since the pACG_XylSPm plasmid series reported unsuccessful induction behavior, the
pACG_araBAD plasmid series were selected for expression regulation. Plasmid pSW3_lacI+ expressing
recombinant mini-proinsulin was additionally transformed into the aforementioned mutants in order
to evaluate effect of genetic regulation on ncBCAA mis-incorporation. At that stage a total of seven
geneX-tunable E. coli strains were engineered, one for each target gene. Each engineered tunable E.
coli strain lacked a certain endogenous target gene, generally indicated as geneX in this work (E. coli
K-12 BW25113 ΔgeneX), and contained one plasmid allowing exogenous expression regulation of
such target gene (pACG_araBAD_geneX) and one plasmid enabling expression of recombinant mini-
proinsulin for subsequent analysis of the impurity profile (pSW3_lacI+). The engineered tunable E. coli
mutants were screened in a mini-reactor by triggering induction of different expression levels of the
target genes. The impurity profile of recombinant mini-proinsulin was then compared with the
control non-engineered E. coli host. Screening was performed in fed-batch mode under standard
cultivation conditions and under conditions mimicking large-scale effects i. e. pyruvate pulsing and O2
limitation. After screening in the mini-reactor system, potential tunable E. coli strains expressing a
certain level of the target gene resulting in an improved impurity profile were scaled-up in a 15L
reactor and impurity profile of the recombinant protein was analysed in order to confirm
reproducible behaviour. Validated E. coli strains reporting a reduced ncBCAA mis-incorporation may
have a crucial role in the pharmaceutical industry since product quality is pivotal for commercial
approval of recombinant proteins that are to be used as human therapeutics (Apostol et al., 1997).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 3
2. Literature Review
2.1 E. coli as industrial biofactory
2.1.1 Advantages of E. coli as industrial biofactory
The Gram-negative bacterium E. coli has been the most extensively used microbial host for
recombinant protein manufacturing in the field of research and pharmaceutical industry, as well as
for metabolic engineering purposes. Somatostatin was the first human recombinant protein to be
expressed in this platform in 1977 (Itakura et al., 1977) while humulin was the first commercially
approved biopharmaceutical produced in E. coli in 1982 (Johnson, 1983). Since then, a total of 45
recombinant drugs produced in E. coli have been commercially approved for human use,
representing about 29.8% of the market of recombinant biopharmaceuticals (Ferrer-Miralles, 2009).
Lantus (Insulin glargin), also manufactured in E. coli, was the 4th top-selling biopharmaceutical
product in 2013, with 7.95 billion $ sales (Walsh, 2014), thus revealing the high impact of E. coli in the
pharmaceutical industry, even nowadays, when the use of non-prokaryotic expression systems is
increasing. There are several advantages making E. coli appealing as a host organism for protein
production purposes. E. coli has been for years extensively characterized in relation to physiology
and genetics, being its genomic sequence available and annotated (Blattner et al., 1997). Its
manipulation by recombinant DNA technology is greatly developed: various molecular techniques are
available for genome editing, hence allowing strain improvement and reverse genetics (Madyagol et
al., 2011; Nakashima and Miyazaki, 2014), and easy protocols for transformation of E. coli with
exogenous plasmid DNA are also well-established, thus facilitating expression of a wide range of
recombinant proteins (Cohen et al., 1972; Hanahan, 1983; Fiedler and Wirth, 1988; Shiloach and
Fass, 2005). E. coli features fast growth kinetics, with a minimal doubling time as short as 20 min
(Sezonov et al., 2007). Moreover, there is a wide range of cultivation procedures with inexpensive
media enabling high cell density cultures leading to high productivity (Riesenberg, 1991; Lee, 1996;
Choi et al., 2006). Furthermore, scalability of E. coli based bioprocesses up to industrial scale is easily
achieved.
2.1.2 Origin and genetic background of strain E. coli K-12 BW25113
E. coli K-12 BW25113 is a frequently used laboratory strain, which was first reported by Datsenko and
Wanner (2000). It is the parent strain for the so-called Keio Collection, which contains around 4000
single gene knock-out mutants (Baba et al., 2006). The complete 4631469-bp genome sequence of
strain E. coli K-12 BW25113 was made available by Grenier et al. (2014) and it was deposited in
GenBank under accession number CP009273. Strain BW25113 was derived from strain BD792 after a
13-step process consisting of serial transduction and allele replacement steps. Strain BD792 was, in
turn, derived from the ancestral E. coli K-12 (EMG2) after a 2-step process. Strain BD792 and its
predecessors contain the rpoS396(Am) allele, which has a base pair substitution that causes an
amber mutation, hence triggering premature translation termination of the stationary-phase sigma
factor σS. However, strain BW25113 carries the pseudo-revertant rpoS(Q33) allele (Hayashi et al.,
2006). In a revertant mutant the original mutation itself is mutated back to the wild type version
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
4 Literature Review
while in a pseudo-revertant mutant the original mutation stays invariable while another mutation
occurring in a different position within the same gene restores the wild type phenotype.
Strain E. coli K-12 BW25113 is characterized by the following genotype: F-, λ-, lacI+, Δ(araD-araB)567,
Δ(rhaD-rhaB)568, ΔlacZ4787(::rrnB-3), rph-1, ilvG-1, hsdR514. It lacks the F-plasmid necessary for
gene transfer and the genes for motility (F-) as well as the lambda lysogen (λ-). This strain was
originally reported to be lacIq (Datsenko and Wanner, 2000), but it was later demonstrated to be lacI+
(Grenier et al. 2014; Baba et al., 2006). araBAD and rhaBAD genetic regions were deleted while a
section of lacZ was replaced with four tandem copies of the rrnB transcriptional terminator
(Datsenko and Wanner, 2000). Deletion of araBAD extends from around 25 bp upstream of the araB
start codon to about 8 bp into the beginning of the araD gene. Hence, the strain is defective in
arabinose (ΔaraBAD), rhamnose (ΔrhaBAD) and lactose (ΔlacZ) metabolizing enzymes, which enables
proper functionality of arabinose, rhamnose and lactose inducible promoters, respectively.
Moreover, frameshift mutations in rph, ilvG and hsdR resulting in a premature translation stop codon
were reported.
The rph gene encodes for an exoribonuclease which plays a crucial role in tRNA maturation (Bowden
et al., 2017) and rRNA degradation (Basturea et al., 2011). rph is organized together with pyrE in the
same operon. pyrE encodes for an orotate phosphoribosyltransferase, which is involved in the
biosynthesis of pyrimidine nucleotides. An intercistronic pyrE attenuator is located downstream of
rph before open reading frame of pyrE (Poulsen et al., 1983). In E. coli K-12 strains, a GC base pair
deletion close to the end of rph genetic region results in a premature translation termination. As a
consequence, rph gene product is truncated (rph-) and lacks exoribonuclease activity. Moreover, the
premature translation termination of rph leads to a deficiency in pyrE expression, since a proper
coupling between rph transcription and translation in necessary in order to maintain optimal
transcription levels of pyrE. E. coli K-12 strains are characterized by pyrimidine starvation when
cultivated in medium lacking exogenous pyrimidine sources such as uracil (Jensen, 1993).
The ilvG gene encodes for acetohydroxy acid synthase II (AHAS II), which is involved in the
biosynthesis of isoleucine starting from α-ketobutyrate. In E. coli K-12 strains, a two-base insertion
event is present in the coding sequence of gene ilvG between base pair positions 1250 and 1253,
resulting in a frameshift mutation. This insertion causes a shift of the reading frame and, as a
consequence, a stop codon is formed, resulting in a premature termination of ilvG gene translation
(ilvG-). AHAS II is hence not expressed and distal expression of ilvEDA operon is impaired. The
absence of AHAS II activity in E. coli K-12 strains leads to the valine toxicity phenomenon: a valine-
mediated inhibition of E. coli growth (Yoon et al., 2012; Parekh and Hatfield, 1997; Lawther et al.
1981). However, frameshift mutations affecting rph and ilvG can be abolished by applying genome
engineering strategies (Hirokawa et al., 2013; Biryukova et al., 2010). Genome sequencing and
annotation studies revealed that, as opposed to E. coli K-12 variants, strain E. coli BL21(DE3) contains
intact ilvG and rph genes. As a consequence, B strains grow better than K-12 variants in minimal
media and growth is not inhibited by valine (Yoon et al., 2012).
2.1.3 Inclusion body-based recombinant protein expression in E. coli
Recombinant proteins expressed in E. coli are generally accumulated in so-called inclusion bodies,
which are highly stable insoluble protein aggregates resistant to proteolysis. Size of IB particles
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 5
ranges from 50 to 800 nm. IBs can be seen as electrodense dark particles polarly distributed in the
cell under transmission electron microscope but they are often even visible in light microscopy (Rinas
et al., 2017). IB formation occurs due to disequilibrium between protein translation, folding and
aggregation rates (Kiefhaber et al., 1991). Protein aggregation mainly happens when folding
intermediates displaying hydrophobic residues at their surface accumulate due to insufficient
presence of chaperones. These folding intermediates are prone to aggregate with each other
through strong intermolecular hydrophobic interactions, which confer high stability to IBs (Mayer
and Buchner, 2004). It was generally considered that IBs comprise non-active protein forms.
Nevertheless, a vast palette of publications reporting biologically active IBs was recently made
available (reviewed by García-Fruitós, 2010). These evidences might benefit the enzymatic industry
since active IBs containing recombinant enzymes could be directly employed as catalyzers without
the necessity of tedious refolding steps (Martínez-Alonso et al., 2009). IBs mainly contain the
recombinant protein, which can represent up to 95 % of the total protein, but also minor cellular
protein rests, phospholipids and nucleic acids, which co-precipitate during the IB isolation process.
Among cellular proteins, membrane proteins are mostly present in the IB fraction, such as proteases
OmpT, OmpF, OmpC and OmpA. In addition, cytoplasmic proteins can also be found in the IB fraction
in lower levels. This is the case of the ribosomal subunit proteins L7/L12, the chaperons DnaK and
GroEL as well as the heat-shock proteins IbpA and IbpB. Plasmid-encoded proteins involved in the
acquirement of antibiotic resistance, such as β-lactamase precursor and kanamycin resistance
protein can also be present in the IB fraction (Jürgen et al., 2010; Rinas and Bailey, 1992; Hartley and
Kane, 1988). Furthermore, the protein composition of IBs is dependent on a number of factors such
as the selected recombinant protein to be expressed, the bacterial host, the cultivation conditions
and the employed IB purification methods (Rinas et al., 2017).
Although E. coli does not have an efficient secretion system to produce recombinant proteins
extracellularly and despite that downstream processing of recombinant proteins present in IBs is
complex, E. coli is still the first choice when expressing simple proteins which do not require
significant post-translational modifications. The main advantage of IB-based recombinant protein
production is that product is resistant to proteolytic degradation and it is obtained at high
concentrations and almost pure. Moreover, physicochemical properties inherent to IBs enable an
easy isolation from the cellular rests by using standard protocols. Additionally, IBs can also be used as
a model for the study of conformational diseases, as drug delivery systems, biomaterials and
immobilized catalysts (Rinas et al., 2017; Ramón et al., 2014). Novel approaches improving recovery
of functional proteins from IBs are reviewed by Ramón et al. (2014) and are mainly based in the
optimization of solubilization and refolding steps. Recent methodologies leading to a reduction of IB
formation during recombinant protein production processes are summarized by Fahnert et al. (2004)
and Zhu et al. (2013), including reduction of translation rate, use of fusion proteins, co-expression of
chaperones and foldases, use of folding promoting agents, oxidizing cytoplasmic redox potential and
alteration of cultivation conditions. Even strategies triggering IB formation, mainly consisting in the
fusion of aggregation-prone peptides to the recombinant protein, are reviewed by Rinas et al. (2017).
2.1.4 Insulin expression in E. coli
Insuline is a peptide hormone which regulates the level of sugar in the blood. It stimulates glucose
uptake by the adipose, muscle and liver tissues, thus having a crucial role in regulating metabolism of
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
6 Literature Review
carbohydrates, fats and proteins. The pre-pro-insulin precursor of human insulin (PPI) is encoded by
gene INS. It is synthesized in the endoplasmic reticulum (ER) of beta cells present in the islets of
Langerhans located in the pancreas. PPI is a biologically inactive form of insulin, containing 110
amino acids organized in A-chain, C-peptide, B-chain and an N-terminal signal peptide, which guides
the nascent protein chain into the ER lumen. The signal peptide is then cleaved by signal peptidases,
yielding pro-insulin (PI). Conditions present in the ER lumen enable folding of PI into the proper
conformation as well as formation of disulfide bonds. PI is then transported through the Golgi
apparatus, where maturation of PI into active insulin takes place. Various peptidases trigger the
cleavage of the C-peptide, which is located between A- and B-chains, yielding insulin. The C-peptide
has a crutial role in the insulin biosynthesis since it joins both A- and B-chains in a manner that allows
proper folding and interchain interaction through disulfide bond formation. The C-peptide itself has
also been reported to activate Ca2+-dependent intracellular signaling pathways upon binding to cell
surface receptors as well as to enhance the Na+-K+-ATPase and the endothelial nitric oxide synthase
activities, among others (Wahren et al., 2000). The active insulin form is then transported by
secretory vesicles through the cell, being finally secreted into the blood circulation. The human
insulin has 51 amino acids, comprising A- and B-polypeptide chains, which remain linked by 2
disulfide bonds. A-chain also contains an intrapeptide disulfind bond (Beals et al., 2008; Zündorf and
Dingermann, 2001).
Since 1920s insulin was extracted from pork and beef pancreas glands collected from rests of the
meat industry. However, insulin from animal sources could trigger immunologic sensitization when
injected in human patients. Moreover, animal sources were not sufficient to supply an increasing
diabetic population and it was then necessary to explore alternative sources of insulin. The successful
expression of human insulin in E. coli by applying the recombinant DNA technology was reported in
1978, ensuring a reliable, safe and constant supply of insulin to diabetic patients around the globe
(Johnson, 1983). Human insulin was the first commercially approved biopharmaceutical produced in
E. coli (1982) and, since then, it has been one of the top-selling biopharmaceutical products on the
market, which is in accordance with the increasing prevalence of diabetes among adults (Walsh,
2014).
When expressing recombinant human insulin in E. coli, a modified gene sequence different from the
original one is used. This altered sequence is adapted to the E. coli codon usage and the genetic
region encoding for the signal peptide is normally substituted by the methionine codon (ATG), which
corresponds to the start codon for protein translation. Pro-insuline gene is then cloned into an
expression plasmid under the control of a strong inducible promoter. Moreover, expression plasmids
normally contain an antibiotic selection marker in order to ensure plasmid stability (Zündorf and
Dingermann 2001). High yield expression of recombinant pro-insulin in E. coli leads to pro-insulin
aggregation and intracellular IB formation due to the lack of chaperones and the high concentration
of folding protein (Mayer and Buchner, 2004). Pro-insulin-containing IBs can represent as much as 20
% of the total E. coli cellular volume (Williams et al. 1982). Upon completion of E. coli cultivation, IBs
have to be isolated from the E. coli cellular rests by cell disruption and then be solubilized in order to
release the free pro-insulin molecules. However, at that stage, recombinant pro-insulin is not active
since the complex mechanism leading to maturation of insulin precursors is not available in E. coli.
Hence, steps conferring the functional peptide conformation taking place in the pancreatic beta cells
have to be reproduced ex vivo: pro-insulin is refolded into the native conformation, the first
methionine residue at N-terminal is removed by CNBr or CNCl treatment, the disulfide bond
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 7
formation is triggered by the so-called oxidative sulfitolysis and the C-peptide is cleaved by treatment
with trypsin and carboxypeptidase B (Figure 1) (Mayer and Buchner, 2004; Vallejo and Rinas, 2004;
Zündorf and Dingermann, 2001).
Figure 1. Expression process of recombinant insulin in E. coli. Adapted from Zündorf and Dingermann (2001).
2.1.5 Physiological stress during recombinant protein expression in E. coli
Recombinant protein production is an artificial stress imposed to the E. coli metabolism, causing
metabolic burden, which refers to the drainage of cell resources (energy molecules such as ATP, GTP
and NAD(P)H; and/or amino acid pools) towards the maintenance and expression of exogenous DNA
within that cell. This metabolic imbalance triggers a vast number of effects: growth arrest due to
reduction in the synthesis of biomass-related proteins, limited recombinant protein biosynthesis,
plasmid instability, accumulation of by-products, ribosome destruction, protein hydrolysis and trigger
of stress responses such as the ppGpp-mediated stringent response (Carneiro et al., 2013). Main
effects of the stringent response comprise inhibition of RNA polymerase and rRNA/tRNA synthesis,
translation repression as well as activation of biosynthetic operons (Lengeler et al., 1999; Traxler et
al., 2008). Metabolic burden reported in E.coli-based recombinant protein production processes
might be similar to the naturally occurring bacteriophage infection, since phages capture the E. coli
metabolic network in order to proliferate and spread infection (Fahnert et al. 2004). However,
available engineering strategies leading to a reduction of the metabolic burden effects are reviewed
by Carneiro et al. (2013).
Misfolded protein variants can accumulate in the E. coli cytosol either by protein aggregation (e.g.
due to recombinant protein production processes) or by protein denaturation (e.g. due to high
temperature). Both cases trigger the induction of the so-called heat shock response. This response
comprises the stabilization and secondary self-induction of the heat-shock sigma factor (σ32) activity,
which, in turn, enhances transcription of numerous heat shock proteins, including Hsp70 and Hsp100
chaperones, Hsp60 chaperonins as well as ATP-dependent Lon and ClpAB proteases. Synthesized
heat shock proteins enable either refolding or degradation of the misfolded protein forms (Fahnert et
al. 2004; Mogk et al., 2015). Likewise, accumulation of misfolded protein variants in the periplasm
induces σ24-dependent envelope shock response (Alba and Gross, 2004).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
8 Literature Review
The general stress response depends on the sigma regulator σs, which activates transcription of
genes inducing transition to stationary phase. General stress response is activated in non-
recombinant E. coli cultivations during glucose starvation and in glucose-limited fed-batch
cultivations (Teich et al. 1999). Moreover, Gill et al. (2000) reported a significant up-regulation of σs
for various E. coli cultivations producing different recombinant proteins. Interestingly, Schweder et
al. (2002) showed that the general stress response is not induced during very strong expression of
recombinant α-glucosidase.
Moreover, Arís et al. (1998) demonstrated that recombinant protein production driven by strong
lambda lytic promoters can promote activation of the SOS response. The SOS proteins trigger DNA
repair and resume cell division once DNA replication is properly restored. Nevertheless, when
metabolic burden associated to recombinant protein expression exceeds a limit, cells are not able to
induce the SOS response, even if DNA damage occurs (Lin et al., 2001).
2.2 Introduction to central E. coli metabolism
2.2.1 Catabolite repression and oxidative respiration
E. coli can metabolize a number of alternative carbon sources in order to obtain energy and
synthesize endogenous constituents for cell growth and maintenance. However, E. coli prefers
glucose than any other carbon source so that, when glucose and other carbon sources are
simultaneously present in the media, E. coli first utilizes glucose and then, the alternative carbon
sources. Glucose preference relies on the catabolite repression mechanism, which ensures
expression inhibition of catabolic operons of carbon sources different than the preferred one.
Glucose depletion (i.e. starvation or limitation) causes de-repression of the catabolic operon of the
next preferred carbon source, hence allowing its utilization.
The catabolite repression mechanism is mainly regulated by 3 enzymes involved in the PEP-
dependent carbohydrate phosphotransferase system (PTS), including sugar-specific PTS permease or
enzyme II (EII), enzyme I (EI) and histidine-containing protein (HPr). These enzymes sequentially
catalyze the transfer of a phosphoryl group from phosphoenolpyruvate (PEP) to the imported sugar,
yielding pyruvate and sugar phosphates. EII comprises 3 protein domains (EIIA, B and C). The
phosphorylated form of EIIA regulates activity of adenylate cyclase, which catalyzes the
transformation of ATP to cAMP and pyrophosphate. In E. coli, catabolite activator protein (CAP) is a
global regulator protein activating the transcription of a variety of genes. cAMP acts as an allosteric
effector and, when bound to CAP, increases CAP’s affinity for DNA, hence allowing CAP’s interaction
with the CAP-binding site, located in the promoter region. CAP can then activate transcription by
interacting with RNA polymerase (Brückner and Titgemeyer, 2002).
The catabolite repression mechanism affecting the lactose operon is presented as the classical
example in E. coli. On the one hand, when glucose is present in the medium, EIIA remains in
unphosphorilated form and adenylate cyclase is inhibited, cAMP levels remain low and CAP does not
bind to the promoter of the lactose operon, hence avoiding lactose consumption. On the other hand,
when glucose is depleted, EIIA switches to the phosphorylated form and adenylate cyclase is
activated, cAMP levels increase and CAP activates transcription of lactose catabolic enzymes.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 9
Glycolysis is the principal pathway for glucose catabolism. It is an oxygen-independent pathway
comprising 10 sequential enzymatic reactions catalyzing conversion of glucose into pyruvate and
release of ATP and NADH molecules. Depending on the environmental conditions, pyruvate can be
further metabolized through two main catabolic pathways: oxidative respiration or fermentation. In
the presence of oxygen, oxidative respiration is activated and pyruvate is decarboxylated to acetyl-
CoA by the pyruvate dehydrogenase complex (PDHC, encoded by genes aceE, aceF and lpdA),
generating CO2 as by-product. Expression of PDHC is down-regulated in the absence of oxygen while
its activity is enhanced by pyruvate (Quail, Haydon, and Guest 1994). Acetyl-CoA can enter the Krebs
cycle, which begins with acetyl-CoA reacting with oxaloacetate to yield citrate and CoA. This reaction
is catalyzed by citrate synthase (CISY, encoded by gene gltA). The Krebs cycle is an 8-step process
catalyzing oxidation of acetyl-CoA to CO2 and release of ATP and NADH. During glycolysis and Krebs
cycle numerous intermediates are formed, which serve as precursors for anabolic pathways. The
generated NADH is then directed to the oxidative phosphorylation pathway. Electrons are
transferred from NADH to oxygen through an electron transport chain, generating ATP. Oxidation of
NADH to NAD+ releases protons into the cytoplasm and energy generated by the electron transport
chain pumps protons across the membrane into the periplasmic space, generating a transmembrane
electrochemical gradient. Protons flow then back across the membrane through ATP synthase,
resulting in ATP generation (Madigan et al. 2014).
2.2.2 Mixed-acid fermentation
In the absence of oxygen as electron acceptor, mixed-acid fermentation is the metabolic pathway
employed for ATP generation in E. coli. However, when it comes to energy production, oxidative
respiration is preferred to mixed-acid fermentation since the first is much more efficient (Sawers et
al., 2004). Mixed-acid fermentation involves the transformation of a hexose into a mix of end
fermentation products, such as acetate, lactate, formate, succinate, CO2 and H2 (Figure 2). All
mentioned fermentation products derive from pyruvate either directly or via acetyl-CoA, with
exception of succinate, which originates from phosphoenolpyruvate via oxaloacetate (Xu et al. 1999).
Each generated fermentation product requires a different set of enzymes to be produced.
Fermentation products are finally secreted extracellularly. The proportion of each end product relies
on a variety of factors, including relative activity of enzymes involved in the fermentation process,
availability of electron acceptors, oxidation form of substrates and presence of redox agents (Liu et
al. 2011). Moreover, bacterial strain can also affect such relative proportion. For instance, acetate is
the dominant product in E. coli K-12 strains and it has been considered one of the major
inconvenients in E. coli-based recombinant protein production processes due to its toxicity (Phue et
al. 2005; Wolfe, 2005).
Under anaerobic conditions, and especially at low pH, lactate dehydrogenase (LDH), encoded by gene
ldhA, catalyzes the conversion of pyruvate into lactate while oxidizing NADH into NAD+, hence
recycling the NADH generated during glycolysis. Formate production is accomplished by pyruvate-
formate lyase (PFL), encoded by gene pflB. PFL catalyzes the transfer of the acetyl group of pyruvate
to CoA yielding formate and acetyl-CoA. Enzyme activity of LDH and PFL is inhibited by the presence
of oxygen and repressed by feedback regulation (Kessler and Knappe 1996). The expression of gene
pflB is regulated by pyruvate accumulation under oxygen limitation conditions (Sirko et al. 1993). PFL
expression is reported as the most sensitive response when modifying oxygen availability in E. coli-
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
10 Literature Review
based cultivations (Schweder et al. 1999). Formate can be further metabolized into CO2 and H2 by the
formate hydrogenlyase complex (FHL), whose functionality strongly relies on the adequate content
of trace elements molybdenum, nickel and selenium in the cultivation medium (Biermann et al.,
2013). FHL complex is a multimeric protein consisting of numerous subunits encoded by genes fdhF
and hycB-G. Mnatsakanyan et al. (2004) showed that FHL plays a key role in H2 and CO2 production at
acidic pH. Acetate synthesis is assisted by the sequential action of two enzymes: phosphate
acetyltransferase (encoded by gene pta) and acetate kinase (encoded by gene ackA). First, phosphate
acetyltransferase reversibly converts acetyl-CoA and phosphate into acetyl-phosphate and CoA,
respectively. Second, acetate kinase reversibly transforms acetyl-phosphate into acetate, while
generating ATP. Pyruvate and PEP play as activators of phosphate acetyltransferase (Campos-
Bermudez et al., 2010). Among all mixed-acid fermentation reactions, the acetate producing pathway
is the only one generating energy in form of ATP. Hence, under anaerobic growth conditions, acetate
kinase is involved in the synthesis of most of the ATP obtained by catabolism. Ethanol formation is
catalyzed by alcohol dehydrogenase (ADH), encoded by gene adhE, in a 2-step reaction. In the first
step acetyl-CoA is converted to acetaldehyde and CoA while oxidizing NADH into NAD+. In the second
step acetaldehyde is converted to ethanol while oxidizing NADH into NAD+. Succinate synthesis
involves action of four different enzymes. First, phosphoenolpyruvate (PEP), a glycolysis
intermediate, is carboxylated by PEP carboxylase (encoded by gene ppc) to yield oxaloacetate.
Second, oxaloacetate is transformed to malate by malate dehydrogenase (encoded by gene mdh),
while oxidizing NADH into NAD+. Third, fumarate hydratase (encoded by gene fumB) catalyzes
dehydration of malate to form fumarate. Finally, fumarate is converted to succinate thanks to
fumarate reductase complex (encoded by genes frdA-D), while oxidizing NADH into NAD+.
2.2.3 Overflow metabolism
Overflow metabolism refers to the phenomenon where incomplete glucose oxidation by
fermentation is favored instead of the more energetically efficient complete oxidation by oxidative
respiration, even under aerobic conditions. When oxygen is present but glucose is in excess in the
cultivation medium, glycolysis and PDHC reactions are stimulated, resulting in high amounts
of ATP and NADH as well as accumulation of pyruvate and acetyl-CoA. This decreases the need for
production of additional energy and reducing equivalents by the Krebs cycle. As a consequence, the
Krebs cycle is blocked and oxygen consumption is reduced. Under this scenario, NAD+ and CoA pools
are significantly drained. However, the glycolytic enzyme glyceraldehyde-3-phosphate
dehydrogenase (GAPDH, encoded by gene gapA) requires NAD+ and PDHC requires CoA as substrate,
respectively. Hence, reoxidation of NADH into NAD+ and recyclying of CoA must be ensured in order
to maintain the glycolytic flux towards pyruvate and acetyl-CoA. Moreover, excess of pyruvate and
acetyl-CoA must be consumed. Since Krebs cycle is inhibited, this is achieved by activating the
reactions of the mixed-acid fermentation (Figure 2). While acetate production serves to generate
ATP, formation of other mixed-acid fermentation products regenerates NAD+ and CoA pools (Wolfe,
2005).
In E. coli acetate is the predominant product of the overflow metabolism (B. Xu et al. 1999) and it is
mainly generated by the sequential action of phosphate acetyltransferase and acetate kinase in the
so-called PTA-ACKA pathway. Alternatively, pyruvate oxidase (encoded by gene poxB) triggers the
direct oxidative decarboxylation of pyruvate into acetate and CO2 under aerobic conditions, while
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 11
reducing FAD to FADH2. The reversible PTA-ACKA pathway can also catabolize acetate, but much less
efficiently than the Acetyl-CoA synthetase pathway (Brown et al., 1977). Acetyl-CoA synthetase (ACS,
encoded by gene acs) converts acetate in acetyl-CoA by consuming ATP and it is subjected to
catabolite repression (Valgepea et al., 2010). Overflow metabolism causes an intracellular
accumulation of pyruvate which, in turn, activates PDHC and PTA (Quail et al., 1994; Campos-
Bermudez et al., 2010), resulting in a significant acetate production. E. coli cells switch from acetate
production (overflow metabolism) to acetate consumption modus mainly by regulating expression
of acs. Activity of ACS is repressed when glucose is in excess while it is induced when glucose is
depleted and acetate is present. Hence, ACS enables the cell to use acetate under aerobic conditions
in order to generate energy via the Krebs cycle, once glucose is depleted (Wolfe, 2005).
Acetate accumulation has proven to be toxic for E. coli cells, having negative consequences at
physiological level. Acetate can easily cross cell membrane, altering the transmembrane pH gradient.
Moreover, it causes cytoplasm acidification, increases osmotic pressure and jeopardizes methionine
biosynthesis (reviewed by Wolfe, 2005). In E. coli-based recombinant protein production processes
acetate accumulation results in a reduction of growth rate and product yields (Eiteman and Altman,
2006).
Furthermore, the intensity of overflow metabolism and the subsequent acetate accumulation is
strain-dependent. E. coli K-12 strains are reported to have a higher acetate accumulation than B
strains. This is commonly explained by three metabolic activities reported to be more active in B
strains: acetate uptake by acetyl-CoA synthetase, glyoxylate shunt encoded by the acetate operon
aceBAK and Krebs cycle (Phue et al. 2005; Yoon et al., 2012).
Several strategies have been made available, leading to a reduction of overflow metabolism and
acetate overproduction in E. coli, including process or strain engineering approaches (reviewed by
Eiteman and Altman, 2006). Recently, Anane et al. (2017) elaborated a model describing overflow
metabolism.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
12 Literature Review
Figure 2. Schematic representation of mixed-acid fermentation and overflow metabolism pathways present in
E. coli. Dashed red and black lines display pyruvate flux towards Krebs cycle. Dashed green lines indicate
acetate synthesis via overflow metabolism. 1: glycolytic enzymes; 2: pyruvate kinase; 3: pyruvate
dehydrogenase complex; 4: citrate synthase; 5: lactate dehydrogenase; 6: pyruvate-formate liase; 7: formate
hydrogenlyase complex; 8: phosphate acetyltransferase; 9: acetate kinase; 10 and 11: alcohol dehydrogenase;
12: phosphoenolpyruvate carboxylase; 13: malate dehydrogenase; 14: fumarate hydratase; 15: fumarate
reductase complex. Genes encoding aofrementioned enzymes appear in italics. Adapted from Reitz (2017).
2.2.4 BCAA biosynthetic pathway
A schematic overview of the BCAA biosynthetic pathway is shown in Figure 3.
2.2.4.1 Biosynthesis of isoleucine and valine
The first step of the isoleucine biosynthesis pathway is catalyzed by threonine deaminase (encoded
by gene ilvA), a pyridoxal 5’-phosphate-dependent enzyme which transforms L-threonine to α-
ketobutyrate and ammonia in a two-step reaction. First, L-threonine is dehydrated to yield
enamine/imine intermediates. Second, intermediates are deaminated to form α-ketobutyrate and
ammonia (Umbarger and Brown, 1957). An alternative source of α-ketobutyrate is the direct carbon
chain elongation of pyruvate catalyzed by enzymes encoded by the leuABCD operon, which
comprises three enzymatic reactions. First, α-isopropylmalate synthase (encoded by gene leuA)
condensates the acetyl group of acetyl-CoA with pyruvate to form citramalate and CoA. Second, α-
isopropylmalate isomerase (encoded by genes leuC-D) catalyzes a 2-step isomerization of citramalate
to β-methyl-D-malate, via citraconate. Third, a NAD-dependent β-isopropylmalate dehydrogenase
Glucose
PEP
Pyruvic acid For mic acid CO2 + H2
Acetyl-CoA
Lactic acid
Acetylaldehyde
Ethanol
Acetyl-Phosphate
Acetic acid
Oxaloacetic acid
Malic acid
Fumaric acid
Succinic acid
Acetyl-CoA
TCA-cycle
CO2
NADH+H+
NAD+
H2O
NADH+H+
NAD+
ADP
ATP
CoA-SH
NAD++CoA-SH
NADH+H++CO2
NADH+H+
NAD+
CoA-SH
Pi
CoA-SH
ADP
ATP
NADH+H+
NAD++CoA-SH
NADH+H+
NAD+
Pi+ADP
CoA-SH+ATP
2 CO2
3 NAD++FAD
3 NADH+H++FADH2
GDP
GTP
CO2
1
2
3
4
6 7
8
9
10
11
12
13
14
15
5
ldhA
pflB fdh, hyo
pflpta
ackA
achE
achE
pyk
aceEF
ppc
mdh
fumB
frd
glk, pgi, pfk
fba, tpi, gap
pgk, gpm,eno
acnB, icd, sucAB
sucCD, sdhABCD
fumB, mdh, gltA
frdA-D
aceE-F, lpdA
gltA
fdhF, hycB-G
adhE
adhE
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 13
(encoded by gene leuB) catalyzes the oxidative decarboxylation of β-methyl-D-malate to produce α-
ketobutyrate (Bogosian et al. 1989).
After formation of α-ketobutyrate, the following four reactions of the isoleucine biosynthesis
pathway are driven by the same enzymes catalyzing the parallel steps of the valine biosynthesis
pathway.
First, acetohydroxy acid synthase (encoded by operons ilvBN, ilvGM and ilvIH) catalyzes a 2-step bi-
substrate enzymatic reaction. In the first reaction step pyruvate binds to AHAS, undergoing
decarboxylation to form an active intermediate. In the second reaction step two alternative
substrates can bind to the pre-formed intermediate to generate two alternative acetohydroxy acids:
a second pyruvate molecule to produce α-acetolactate as end-product or α-ketobutyrate to generate
α-acetohydroxybutyrate. α-acetolactate serves as precursor for valine biosynthesis while α-
acetohydroxybutyrate is a precursor of isoleucine. E. coli encodes three AHAS isozymes which
diverge in their biochemical properties. Enzymes AHAS I, II and III are encoded by operons ilvBN,
ilvGM and ilvIH, respectively. AHAS I has substrate preference for pyruvate in the second reaction
step while AHAS II and III have a substrate preference for α-ketobutyrate, being this higher for AHAS
II. Hence, AHAS I directs the metabolic flux to the production of α-acetolactate while AHAS II and III
preferably to the formation of α-acetohydroxybutyrate. In addition, catalytic efficiency (given by the
kcat/Km ratio) of AHAS I is 2-fold and 40-fold higher than AHAS II and AHAS III, respectively (Barak et
al., 1987; Salmon et al., 2006; Vinogradow et al., 2006).
Second, acetohydroxy acid isomeroreductase (encoded by gene ilvC) catalyzes the transfer of the
ethyl group of α-acetohydroxybutyrate or the methyl group of α-acetolactate from the α-carbon to
the β-carbon to form the α,β-dihydroxy acids α,β-dihydroxy-β-methylvalerate and α,β-
dihydroxyisovalerate, respectively.
Third, dihydroxyacid dehydratase (encoded by gene ilvD) catalyzes dehydratation of the α,β-
dihydroxy acid intermediates to yield α-keto acids α-ketoisovalerate (valine precursor) and α,β-keto-
β-methylvalerate (isoleucine precursor).
Fourth, the branched-chain amino acid aminotransferase or transaminase B (encoded by gene ilvE)
catalyzes transamination of α,β-dihydroxy acids by using glutamate as donor of the amino group.
Hence, α-ketoisovalerate is converted into valine while α,β-keto-β-methylvalerate is transformed
into isoleucine. Moreover, valine can also undergo transamination by transaminase C (encoded by
gene avtA) by using alanine as donor of the amino group (Whalen et al., 1982).
2.2.4.2 Biosynthesis of leucine
Leucine biosynthesis comprises the carbon chain elongation of α-ketoisovalerate is catalyzed by
enzymes encoded by the leuABCD operon, which comprises three enzymatic reactions. First, α-
isopropylmalate synthase condensates the acetyl group of acetyl-CoA with α-ketoisovalerate to yield
α-isopropylmalate and CoA. Second, α-isopropylmalate isomerase catalyzes a 2-step isomerization of
α-isopropylmalate to β-isopropylmalate, via dimethyl-citraconate. Third, the NAD-dependent β-
isopropylmalate dehydrogenase catalyzes the oxidative decarboxylation of β-isopropylmalate to
produce α-ketoisocaproate. Transaminase B can then catalyze transamination of α-ketoisocaproate
by using glutamate as donor of the amino group. Hence, α-ketoisocaproate is converted into leucine.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
14 Literature Review
Furthermore, α-ketoisocaproate can also be subjected to transamination by the aromatic amino acid
aminotransferase (encoded by gene tyrB) by using also glutamate as donor of the amino group
(Vartak et al., 1991; Huang et al., 2009).
Figure 3. Schematic representation of the BCAA biosynthetic pathway. See text for description. Dashed lines
indicated BCAA-mediated feedback regulation. AHAS: acetohydroxy acid synthase, encoded by genes ilvBN,
ilvGM and ilvIH; DH: dihydroxy-acid dehydratase, encoded by ilvD; PMS: α-isopropylmalate synthase, encoded
by leuA; IR: ketol-acid reductoisomerase (NADP(+)), encoded by ilvC; IPMD: 3-isopropylmalate dehydrogenase;
encoded by leuB; ISOM: 3-isopropylmalate dehydratase, encoded by leuCD; TD: threonine deaminase, encoded
by ilvA; TrB: transaminase B, encoded by ilvE; AK: bifunctional aspartokinase/homoserine dehydrogenase 1,
encoded by thrA; ASAD: aspartate-semialdehyde dehydrogenase, encoded by gene asd; HSAT: homoserine
acyltransferase; HSD: bifunctional aspartokinase/homoserine dehydrogenase 1, encoded by thrA; HSK:
homoserine kinase, encoded by thrB; TS: threonine synthase, encoded by thrC; TrC: transaminase C, encoded
by avt; tyrB: gene encoding aromatic amino acid aminotransferase. Adapted from Reitz (2017).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 15
2.2.5 Regulation of the BCAA biosynthetic pathway
The branched chain amino acid biosynthesis pathway is effectively modulated by a complex
regulatory network. This enables cells to adapt BCAA biosynthesis to shifting environmental
conditions. Regulation can take place at enzymatic or genetic level. Genetic regulation is hierarchical
and its range of action can be either global or affect local pathways. A summary of the main
regulation mechanisms affecting the target genes investigated in this thesis is provided in Table 1.
Table 1. Summary of the regulation mechanisms affecting the target genes investigated in this thesis.
Gene
Enzyme
Transciptional regulation
Posttranslational
regulation
leuA
2-isopropylmalate
synthase
-Attenuation (by leu)
-Feed-back inhibition (by
leu)
-Lrp regulation (leu does not show a
modulation effect)
-RelA/SpoT modulon (up-regulation by
(p)ppGpp after amino acid starvation)
ilvC
Ketol-acid
reductoisomerase
(NADP(+))
-Substrate-mediated activation (substrate
binding to a preformed IlvY protein DNA
complex relaxes an IlvY protein-induced
DNA bend and increases the affinity for RNA
polymerase)
-
-RelA/SpoT modulon (up-regulation by
(p)ppGpp after amino acid starvation)
ilvIH
acetohydroxyacid
synthase isozyme
III
-Lrp regulation (leu inhibits Lrp activator
from binding to the promoter regulatory
region of ilvIH operon thus decreasing
transcription)
-Feed-back inhibition (by
val, leu and ile)
-RelA/SpoT modulon (down-regulation by
(p)ppGpp after amino acid starvation)
ilvGM
acetohydroxyacid
synthase isozyme II
-Attenuation (val, leu and ile)
-
-Lrp regulation (leu inhibits Lrp repressor
from binding to the transcription site of
ilvGMEDA operon thus increasing
transcription)
-IHF-mediated global regulation (activation)
-RelA/SpoT modulon (up-regulation by
(p)ppGpp after amino acid starvation)
ilvBN
acetohydroxyacid
synthase isozyme I
-Attenuation (by val and leu)
-Feed-back inhibition (by
val)
-CAP-mediated global regulation (activation
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
16 Literature Review
when glc levels are low)
-IHF-mediated global regulation (activation)
-RelA/SpoT modulon (up-regulation by
(p)ppGpp after amino acid starvation)
thrA
threonine-
sensitive,
bifunctional
aspartokinase/
homoserine
dehydrogenase 1
-Attenuation (thr and ile)
-Feed-back inhibition
(thr).
-Activation by ile and
met.
-RelA/SpoT modulon (up-regulation by
(p)ppGpp after amino acid starvation)
ilvA
L-threonine
dehydratase
-
-Feed-back inhibition
(ile). Activation by valine.
-
2.2.5.1 Posttranslational regulation
Allosteric regulation is an important mechanism of posttranslational regulation, which enables cells
to instantaneously adjust their metabolism in response to environmental changes. The activity of the
so-called allosteric enzymes can be adjusted by the non-covalent interaction of effector molecules to
regulatory sites located in the enzyme outside the active site. Binding of the effector (activator or
inhibitor) to the regulatory site triggers conformational changes which might either increase or
decrease enzymatic activity. Allosteric regulation is typical of biosynthetic enzymes. In this case, the
end-product of the biosynthetic pathway normally acts as inhibitor of the enzyme involved in that
pathway. This regulation mechanism is known as feedback inhibition and it usually affects the first
enzyme of a certain biosynthetic pathway. However, allosteric activation of a biosynthetic enzyme is
also not rare (Lengeler et al., 1999). Accordingly to aforementioned, various enzymes involved in the
BCAA biosynthetic pathway are subjected to feedback inhibition and allosteric activation (Table 2).
Table 2. Enzymes involved in the BCAA biosynthetic pathway subjected to allosteric regulation and their
corresponding inhibitors and activators.
Enzyme
Coding
gene
Inhibitor
Activator
Reference
Aspartokinase/homoserine
dehydrogenase
thrA
threonine
Isoleucine,
methionine
Patte (1996)
Threonine deaminase
ilvA
isoleucine
valine
Umbarger (1956),
Monod et al. (1965)
α-isopropylmalate synthase
leuA
leucine
-
Soper et al. (1976)
Acetohydroxy acid synthase
ilvIH
valine, leucine,
-
Salmon et al.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 17
isozyme III
isoleucine
(2006)
Acetohydroxy acid synthase
isozyme I
ilvBN
valine
-
Salmon et al.
(2006)
2.2.5.2 Genetic regulation
In addition to posttranslational regulation, e.g. regulation of enzymatic activity, gene expression
patterns are also strictly regulated in order to ensure adaptation of cells to nutritional and
environmental changing conditions. As opposed to enzyme regulation, genetic regulation reports a
lag time until the response mechanism to changes is effective but it also ensures a long-term
response. Genetic regulation occurs at different hierarchical levels: local, regional and global. Local
genetic regulation affects only a single gene or operon while regional regulation acts on multiple
operons belonging to the same regulon. In an operon, genes coding for similar functions are
organized as a single transcriptional unit (e.g. genes involved in lactose catabolism). A regulon
comprises a set of operons which are dispersed in the genome but are under the control of the same
specific regulator (e.g. operons involved in the biosynthesis of leucine). The integration of genes in
operons and regulons enables an efficient and coordinated regulation of functionally related genes.
Global genetic regulation affects the expression of modulones, which integrate genes, operons
and/or regulons having a common global purpose (e.g. genes involved in anaerobic respiration)
(Lengeler et al., 1999). Genetic units belonging to a modulon are regulated by the same global
regulator. Moreover, a stimulon is a group of genes responding to the same environmental stimulus
(e.g. temperature).
A total of 15 genes involved in BCAA biosynthesis are grouped in 5 operons (ilvBN, ilvGMEDA, ilvIH,
ilvYC and leuABCD), which, in turn, are integrated in the so-called ilv regulon. Transcription of
operons ilvBN, ilvGMEDA, leuABCD and thrABC is regulated by attenuation (Vitreschak et al., 2006).
Operon ilvYC is regulated by an operon-specific mechanism mediated by protein IlvY. In addition,
operons ilvIH, ilvGMEDA and leuABCD are regulated by the global Lrp (leucine-responsive protein).
ilvBN operon is subjected to global catabolite repression control mediated by CAP (catabolite
activator protein) as well. Moreover, the global IHF (integration host factor) also regulates expression
of operons ilvBN and ilvGMEDA (Salmon et al., 2006). Additionally, the global RelA/SpoT modulon is
also described to regulate transcription of operons implicated in amino acid biosynthesis (Fang and
Bauer, 2018).
2.2.5.2.1 Transcription attenuation
Transcription attenuation is a genetic regulation mechanism based in the synchronization of
transcription and translation processes. The nascent RNA transcript of an operon contains a leader
region located between the transcription initiation site and the first coding gene of the operon. The
sequence of the leader region normally contains codons corresponding to amino acids synthesized by
the gene products of the operon, i.e. regulatory codons. Shortly after transcription begins, a 1:2
stem-loop secondary structure is assembled, resulting in the obstruction of RNA polymerase to
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
18 Literature Review
proceed with the transcription process. RNA polymerase remains then momentarily paused at the
end of stem-loop 1:2. This interruption enables the ribosome to start translating the leader peptide.
When ribosome reaches steam 1, RNA polymerase resumes transcription and, from then on, both
transcription and translation processes are coupled. Depending on the availability of certain amino-
acylated tRNAs in the cell, two alternative scenarios can occur next. If the cellular level of necessary
amino-acylated tRNAs is low, ribosome stops at regulatory codons while RNA polymerase continues
transcribing, causing a desynchronization. This scenario allows formation of the anti-terminator 2:3
stem-loop secondary structure, hence impeding formation of transcription termination and enabling
read-through into the structural genes of the operon. However, if the cellular level of necessary
amino-acylated tRNAs is high, ribosome translates the regulatory codons and proceeds until the stop
codon, thus translating the complete leader peptide. This scenario ensures synchronization of
transcription and translation which avoids formation of the anti-terminator 2:3 stem-loop but
triggers assembly of the terminator 3:4 stem-loop secondary structure, i.e. attenuator, resulting in a
premature transcription interruption, i.e. attenuation (Vitreschak et al., 2006; Salmon et al., 2006;
Lengeler et al., 1999). Accordingly, ribosome is the main mediator of the transcription attenuation
mechanism and the degree of attenuation relies on the position of the RNA polymerase when the
ribosome dissociates from the RNA transcript and on the formation probabilities of each RNA
secondary structure (Roesser et al., 1989).
Transcription attenuation of an operon involved in the biosynthetic pathway of a certain amino acid
(e.g. leuABCD operon) is usually mediated by the cellular levels of such amino acid, since the
regulatory codons present in the leader region are codons for such amino acid. However,
transcription attenuation of operons implicated in a biosynthetic pathway shared by various amino
acids (e.g. ilvBN operon) might be mediated by all those amino acids, as the leader region would
contain codons for such amino acids (Vitreschak et al., 2006).
Transcription of operons ilvBN, ilvGMEDA, leuABCD and thrABC is regulated by attenuation. Table 3
shows the regulatory codons present in the leader peptide of the aforementioned operons.
Table 3. Regulatory codons present in the leader peptide of various operons involved in the BCAA biosynthetic
pathway. Data obtained from Lengeler et al. (1999).
Operon
Regulatory codons
Amino acid sequence leader peptide
ilvBN
valine, leucine
MTTSMLNAKLLPTAPSAAVVVVRVVVVVGNAP
ilvGMEDA
valine, leucine, isoleucine
MTALLRVISLVVISVVVIIIPPCGAALGRGLA
leuABCD
leucine
MSHIVRFTGLLLLNAFIVRGRPVGGI
thrABC
threonine, isoleucine
MKRISTTITTTITITTGNGAG
2.2.5.2.2 IlvY-mediated regulation of ilvYC operon
As opposed to aforementioned operons, which are regulated by the end product of the specific
biosynthetic pathway they are involved in, operon ilvYC is regulated by its substrates, e.g. α-
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 19
acetolactate. Genes ilvY and ilvC are organized in the same operon in opposite orientations and their
transcription is controlled by the divergent overlapping ilvYC promoter. IlvY is a regulatory protein
encoded by gene ilvY that can bind to operator sites O1 and O2, located in the promoter region. When
no substrates of α-acetohydroxy acid isomeroreductase (encoded by ilvC) are present, ilvY strongly
binds to the operator regions, forming a DNA bend that inactives ilvYC promoter. When substrates
are present, they can bind to the pre-formed IlvY-operator complex, resulting in a relaxation of the
DNA bend and a dramatic increase of RNA polymerase affinity for ilvC promoter (Arfin et al., 1969;
Rhee et al., 1998).
2.2.5.2.3 Global regulation
2.2.5.2.3.1 Lrp-mediated global regulation
The leucine-responsive regulatory protein (Lrp) is a global regulator of the transcription process. It
regulates about 10 % of the E. coli genes, being most of them involved in the transport and
metabolism of amino acids. Depending on the target operon, expression can be either activated or
repressed upon Lrp action. Lrp recognizes and binds to DNA target sites generally located in the
promoter region of the target operon. Leucine usually acts as an effector ligand of Lrp, being
sometimes required for Lrp binding to target sites. However, in other cases, leucine can also impede
or have no effect on the interaction of Lrp with the DNA target sites. In general, Lrp promotes
expression of biosynthetic operons while repression of catabolic ones (Rhee et al., 1996; de los Rios
and Perona, 2006).
Operons ilvIH, ilvGMEDA and leuABCD were demonstrated to be regulated by the global Lrp protein
in different manners (Lin et al., 1992; Platko et al., 1990; Rhee et al., 1996). Lrp directly triggers
activation of the ilvIH promoter and addition of leucine into the cultivation medium reduces binding
capacity of Lrp to the target sites located at the ilvIH promoter region, resulting in a 5- to 10-fold
transcription decrease (Chen et al., 2005; DeFelice and Levinthal, 1977). As opposed to Lrp-mediated
regulation of ilvIH operon, Rhee et al. (1996) showed that Lrp triggers transcriptional repression of
the ilvGMEDA operon. Lrp binds to DNA in a position between the leader region and the first
structural gene of the operon (ilvG). Moreover, leucine reduces binding capacity of Lrp to the target
sites of the ilvGMEDA operon, resulting in a decrease of the repression degree. Lin et al. (1992) and
Tchetina and Newman (1995) demonstrated that Lrp regulates expression of leuABCD operon.
Different from aforementioned operons, leucine does not show a modulation effect on the Lrp-
mediated regulation of leuABCD expression (Lin et al. 1992).
The monomer form of Lrp contains two protein domains: an N-terminal helix-turn-helix motif and a
C-terminal antiparallel β-sheet flanked by two α-helices. The N-terminal domain allows binding of Lrp
to DNA while the C-terminal domain contains a leucine-binding site. Binding of leucine to Lrp triggers
a conformational alteration of Lrp, resulting in an expression regulation of the target genes through
modulation of Lrp capacity to bind DNA target sites (de los Rios and Perona, 2007).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
20 Literature Review
Expression of the lrp gene can be regulated by Lrp itself by autogenous repression (Wang et al.,
1994). Expression of lrp was also demonstrated to be dependent on the nutrients present in the cell
medium: lrp is down-regulated in rich media, in glucose minimal media supplemented with amino
acids as well as in minimal media containing alternative carbon sources, such as rhamnose,
arabinose, glycerol, pyruvate, acetate or succinate (Chen et al., 1997).
2.2.5.2.3.2 CAP-mediated global regulation
In addition to regulation by transcription attenuation, the ilvBN operon is subjected to global
catabolite repression control mediated by CAP (catabolite activator protein). Freundlich (1977)
suggested that the ppGpp requirement for ilvBN expression could be substituted by cAMP. Sutton
and Freundlich (1980) demonstrated then that cAMP was indeed involved in the regulation of ilvBN
operon, but not ilvIH and ilvGM, since cAMP allows de-repression of AHAS during valine and leucine
but not isoleucine starvation. In presence of glucose (i.e. when cAMP levels are low) activity of AHAS
I remains low. However, when addying cAMP in the presence of glucose or when using carbon
sources different from glucose (i.e. glycerol, succinate or lactate), a 2 to 3-fold increase of AHAS I
activity is reported (Sutton and Freundlich, 1980). Friden et al. (1982) determined the sequence of
the promoter region of the ilvBN operon and shown that regions -35 and -72 have similarities to
regulatory regions present in CRP (cAMP receptor protein)-dependent promoters. Friden et al. (1984,
I) reported that CRP binds to the ilvBN promoter at a position located between -44 to -82 and
prevents RNA polymerase interaction with a second non-productive binding site. Dailey and Cronan
(1986) proposed that CRP-mediated activation of operon ilvBN might allow sufficient isoleucine and
valine biosynthesis when E. coli grows in the presence of a carbon source different from glucose (i.e.
acetate or oleate).
2.2.5.2.3.3 IHF-mediated global regulation
The global IHF (integration host factor) also regulates expression of operons ilvBN and ilvGMEDA. IHF
is a DNA-binding heterodimer encoded by genes ihfA and ihfB and it plays an important role in the
integration of bacteriophage lambda DNA into the E. coli genome, plasmid maintenance, conjugation
and gene expression regulation, among others (Miller et al., 1979; Friedman, 1988). Friden et al.
(1984, II) suggested that IHF might be a positive effector for transcription of operons ilvBN and
ilvGMEDA since IHF E. coli mutants reported decreased amounts of enzymes encoded by those
operons, resulting in severe growth inhibition in minimal medium. IHF-binding sites are located at -85
in the ilvBN operon and at -95 and +131 in the ilvGMEDA operon (reviewed by Freundlich et al.,
1992). IHF interaction triggers expression activation of both ilvBN and ilvGMEDA operons. Tsui and
Freundlich (1990) demonstrated that IHF is a direct activator of ilvBN transcription. When bound to
ilvBN promoter-leader DNA region, IHF triggers DNA bending, thus acting as an antiterminator. This
impedes formation of transcription termination and enables read-through of the RNA polymerase
into the structural genes of operon ilvBN. The promoter of operon ilvGMEDA comprises two
upstream activating sequences: UAS1 and UAS2. UAS1 contains the IHF-binding site. Binding of IHF to
this site when DNA template is negatively supercoiled activates transcription of the ilvGMEDA
operon a maximum of 5-fold (Parekh et al., 1996).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 21
2.2.5.2.3.4 Global RelA/SpoT modulon
Additionally, the global RelA/SpoT modulon is also described to regulate transcription of operons
implicated in amino acid biosynthesis. Under conditions of amino acid or primary carbon source
starvation, E. coli cells trigger a stringent response. This response is characterized by a sudden
accumulation of GTP derivatives guanosine 5’-diphosphate, 3’-diphosphate (ppGpp) and guanosine
5’-triphosphate, 3’-diphosphate (pppGpp), also known as alarmones. The synthesis of (p)ppGpp is
enhanced during amino acid starvation, when uncharged tRNAs bind to the acceptor site of a
ribosome. The resulting stalled ribosomes activate (p)ppGpp synthetase I (RelA), which is encoded by
gene relA. RelA remains bound to the 50S large ribosomal subunit through peptide L11 (Loveland et
al., 2016) and catalyzes transference of a pyrophosphoryl group from ATP to GTP to yield (p)ppGpp.
An alternative RelA-independendent pathway for (p)ppGpp biosynthesis is activated during long-
lasting carbon starvation and requires SpoT. SpoT, encoded by gene spoT, is a bifunctional (p)ppGpp
synthetase II/hydrolase, regulating (p)ppGpp synthesis and degradation, respectively. Alarmone
levels are mainly controlled by the protein pair RelA/SpoT (termed RSH) but other enzymes such as
pppGpp-5’-phosphohydrolase (Gpp, encoded by gene gppA) and nucleoside diphosphate kinase
(Ndk, encoded by gene ndk) are also involved (Hauryliuk et al., 2015; Lengeler et al., 1999) (Figure 4).
Figure 4. Main synthesis and degradation pathways of alarmone (p)ppGpp in E. coli. See text for description.
Adapted from Lengeler et al. (1999).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
22 Literature Review
In addition, Fang and Bauer (2018) recently found that the ACT domain present in RSH proteins from
numerous bacterial species binds BCAA. This interaction results in a modulation of the alarmone
synthetase/hydrolase activities, thus adding another layer of control of the stringent response. Fang
and Bauer (2018) demonstrated that binding of amino acids valine and isoleucine to the ACT domain
of RSH increases (p)ppGpp hydrolase activity. Accordingly, when BCAA are depleted (p)ppGpp
hydrolysis would remain inhibited, thus triggering accumulation of (p)ppGpp and the subsequent
stringent response.
Main effects of the (p)ppGpp-mediated stringent response comprise inhibition of RNA polymerase
and rRNA/tRNA synthesis, translation repression as well as activation of biosynthetic operons,
including the BCAA biosynthetic enzymes (Lengeler et al., 1999; Traxler et al., 2008). Artsimovitch et
al. (2004) reported that ppGpp controls gene expression by binding to a single site of the RNA
polymerase (RNAP) in two alternative orientations. Moreover, they suggested that this binding may
alter structure of the RNAP active site, hence affecting NTP (nucleoside triphosphate) interactions
with RNAP, and that ppGpp might also interact with the coding DNA strand in the transcription
bubble.
Traxler et al. (2008) demonstrated that genes thrA, thrB and ilvC as well as operons leuABCD and
ilvGMEDA were significantly up-regulated under conditions of isoleucine depletion by the (p)ppGpp-
mediated stringent response. Interestingly, as opposed to ilvGM, operons ilvIH and ilvBN did not
report any significant expression modulation in that study. Other reports show that ilvBN operon is
stimulated by ppGpp (Umbarger, 1996; Tedin and Norel, 2001) while ilvIH expression is actually
repressed by ppGpp under conditions of amino acid starvation (Baccigalupi et al., 1995).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 23
2.3 Non-canonical branched chain amino acids (ncBCAA)
2.3.1 Introduction to ncBCAA
A branched-chain amino acid (BCAA) is an amino acid containing an aliphatic side-chain. The
proteinogenic or canonical BCAA are leucine, isoleucine and valine. The most prominent non-
proteinogenic or non-canonical BCAA include norleucine, norvaline, homoisoleucine and β-
methylnorleucine. The structure of the different BCAA is shown in Table 4.
Table 4. Molecular structure of canonical and non-canonical BCAA
Canonical BCAA
Non-canonical BCAA
Amino acid
Molecular structure
Amino acid
Molecular structure
Leucine
Norleucine
Norvaline
Isoleucine
Homoisoleucine
Valine
β-methylnorleucine
The non-canonical BCAA can be produced by the E. coli metabolism as byproducts of the BCAA
biosynthetic pathway under particular conditions. These modified amino acids can be secreted to the
medium and/or mis-incorporated into cellular proteins through tRNA misaminoacylation during
protein translation. Such mis-incorporation can lead to the production of altered proteins, having
non optimal characteristics e.g. altered biological activity, modulated sensitivity to proteolysis and
immunogenicity (Laird and Veeravalli, 2013). Hence, it has become a crucial matter of contention in
the pharmaceutical industry since product quality (high purity and homogeneity) is pivotal for
commercial approval of recombinant proteins that are to be used as human therapeutics (Apostol et
al., 1997). However, non-canonical amino acids, including ncBCAA, also awakens interest in the field
of synthetic biology since directed mis-incorporation of these translationally active analogues opens
the possibility to develop proteins with new functionalities. The main strategy employed to trigger
mis-incorporation of a given non-canonical amino acid in positions corresponding to the canonical
OH
O
NH2
CH3
CH3
OH
O
NH2
CH3
OH
O
NH2
CH3
OH
O
NH2
CH3
CH3
CH3OH
O
NH2
CH3
CH3OH
O
NH2
CH3
OH
O
NH2
CH3
CH3
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
24 Literature Review
counterpart (residue-specific incorporation of non-canonical amino acids) is the supplementation of
that non-canonical amino acid into the cultivation medium while using an E. coli strain auxotrophic
for the canonical counterpart. Under those conditions, E. coli is forced to utilize the supplemented
non-canonical amino acid for protein translation (Kiick et al., 2001). This strategy allows then the
global replacement of a given residue by a non-canonical counterpart. However, there are also
strategies enabling replacement of only a single residue in a certain position of the protein (site-
specific incorporation of non-canonical amino acids). The traditional strategy consists of the
supplementation of chemically aminoacylated suppressor tRNA with the non-canonical amino acid to
an in vitro cell-free system where target codon of the gene encoding the recombinant protein is
substituted by an amber nonsense codon (TAG) (reviewed by Link et al., 2003). Recently, a new
strategy was developed enabling incorporation of (synthetic) non-canonical amino acids into
recombinant proteins in living cell systems. This includes the transference of an orthogonal
translation system including the aminoacyl-tRNA synthetase (aaRS) and the tRNA pair corresponding
to the non-canonical amino acid of interest as well as substitution of the target codon by a dedicated
one (reviewed by Young and Schultz, 2010).
Among the non-canonical BCAA, norleucine is the one which accumulates more literature
knowledge. The first evidence of the incorporation of exogenous norleucine into a recombinant
protein by E. coli dates from 1956 (Munier and Cohen, 1956). Then, different studies reported that
the mis-incorporation of exogenous norleucine into recombinant proteins by E. coli took place at
positions where methionine is normally incorporated (Cohen and Munier, 1959; Cowie et al., 1959),
thus confirming that norleucine is a structural analog of methionine. From that period onwards
plenty of literature was made available demonstrating exogenous norleucine mis-incorporation into
a wide range of recombinant proteins by the E. coli production platform in methionine positions,
including recombinant adenylate kinase (Gilles et al., 1988), recombinant mammalian calmodulin
(Yuan and Vogel, 1999) and recombinant cytochrome P450 BM-3 heme domain (Cirino et al., 2003).
There were also cases reported where norleucine was not being supplied exogenously in the media,
but was being naturally synthetized in E. coli cells and incorporated into recombinant proteins. In
that case norleucine was found to be incorporated into recombinant interleukin-2 (IL-2) (Lu et al.,
1988), recombinant bovine somatotropine (bST) (Bogosian et al., 1989), recombinant human
macrophage colony stimulating factor (hM-CSF) (Randhawa et al., 1994), recombinant human brain-
derived neurotrophic factor (Sunasara et al., 1999) and in a 41 kDa Met-rich recombinant protein
vaccine candidate (Ni et al., 2015). Despite all cases described for recombinant proteins, there are no
evidences in the literature regarding the synthesis and incorporation of norleucine into natural non-
recombinant proteins by E. coli. However, the natural presence of norleucine has been reported in
the field-growing parasitic fungi Claviceps purpurea (Cvak et al., 2005).
Formation of norleucine by complex regulatory mutants of Serratia marcescens has been reported
(Kisumi et al. 1976, I and II). The expression of recombinant leucine-rich proteins may trigger the
biosynthesis of norleucine in E. coli (Bogosian et al., 1989). Norleucine can be incorporated into
proteins both at internal residues as well as the amino terminus and the incorporation at the
methionine loci in the protein was demonstrated to be random (Bogosian et al., 1989). However,
recent evidences reveal that, at certain positions, higher levels of norleucine incorporation are
reported (Veeravalli et al., 2015).
Norvaline was initially reported as a natural component of an antifungal peptide manufactured in
Bacillus subtilis (Nandi and Sen, 1953). Norvaline was identified as a byproduct in isoleucine
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 25
overproducing regulatory mutants of Serratia marcescens (Kisumi et al., 1976, I and II). Similarly to
norleucine, the incorporation of norvaline into recombinant proteins at leucine positions can be
intentionally triggered by feeding microorganisms with exogenous norvaline (Miyazawa et al., 1989).
In addition, norvaline can also be naturally synthetized in E. coli cells and incorporated into
recombinant proteins (recombinant hemoglobin, Apostol et al., 1997). Recent experimental results
have shown that the in vivo production of norvaline by E. coli is boosted in conditions of oxygen
limitation (Soini et al., 2008, I).
Additionally, other reports indicate that the non-canonical amino acids β-methylnorleucine
(recombinant hirudin, Muramatsu et al., 2002) and homoisoleucine (recombinant human brain-
derived neurotrophic factor, Sunasara et al., 1999) are sometimes inappropriately inserted into
heterologous proteins, in the place of isoleucine. β-methylnorleucine was first discovered to be
produced by Serratia marcescens (Sugiura et al., 1981). Biosynthesis of homoisoleucine in Serratia
marcescens is described in Kisumi et al. (1976, I and II). The incorporation of homoisoleucine into the
recombinant coiled-coil peptide A1 was induced by homoisoleucine addition to the medium by Van
Deventer et al. (2011).
2.3.2 ncBCAA biosynthetic pathway
The initial outline for norleucine, norvaline and homoisoleucine biosynthesis was proposed by Kisumi
et al. (1976, I and II) by using regulatory mutants of leucine biosynthesis in Serratia marcescens. The
scheme for β-methylnorleucine biosynthesis was later suggested by Sugiura et al. (1981). Some
evidences also indicate that E. coli utilize the same metabolic pathway than Serratia marcescens to
synthesize non-canonical branched-chain amino acids (Tsai et al., 1988; Bogosian et al., 1989;
Muramatsu et al., 2003), confirming that it derives from the isoleucine route.
A schematic representation of the ncBCAA biosynthetic pathway is shown in Figure 5. The first
substrate of both isoleucine and ncBCAA biosynthetic pathways is α-ketobutyrate, which can be
mainly synthesized from the deamination of threonine by threonine deaminase (ilvA) but also from
pyruvate via keto acid chain elongation by the leucine operon (leuABCD) (Bogosian et al., 1989). α-
ketobutyrate can be either transformed to isoleucine by ilv operons or sequencially converted to α-
ketovalerate via keto acid chain elongation by the leucine operon (leuABCD). In detail, α-
isopropylmalate synthase condensates the acetyl group of acetyl-CoA with α-ketobutyrate to
produce α-ethylmalate and CoA. Then, α-isopropylmalate isomerase catalyzes isomerization of α-
ethylmalate to β-ethylmalate. Later, the NAD-dependent β-isopropylmalate dehydrogenase catalyzes
the oxidative decarboxylation of β-ethylmalate to produce α-ketovalerate. α-ketovalerate can, in
turn, be either transformed to α-keto-β-methylcaproate by ilv operons, either sequencially converted
to α-ketocaproate via keto acid chain elongation by the leucine operon (leuABCD) or subjected to
transamination by transaminase B to yield norvaline (Tsai et al., 1988; Bogosian et al., 1989; Sycheva
et al., 2007; Soini et al., 2008, I).
Keto acid chain elongation of α-ketovalerate by the leucine operon (leuABCD) is described in detail
below. α-isopropylmalate synthase condensates the acetyl group of acetyl-CoA with α-ketovalerate
to produce α-propylmalate and CoA. Then, α-isopropylmalate isomerase catalyzes isomerization of
α-propylmalate to β-propylmalate. Later, the NAD-dependent β-isopropylmalate dehydrogenase
catalyzes the oxidative decarboxylation of β-propylmalate to produce α-ketocaproate. Transaminase
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
26 Literature Review
B catalyzes then transamination of α-ketocaproate to yield norleucine (Tsai et al., 1988; Bogosian et
al., 1989; Sycheva et al., 2007; Soini et al., 2008, I).
Transformation of α-ketovalerate by ilv operons is shown in detail next. First, acetohydroxy acid
synthase catalyzes a 2-step bi-substrate enzymatic reaction. In the first reaction step pyruvate binds
to AHAS, undergoing decarboxylation to form an active intermediate. In the second reaction step α-
ketovalerate binds to the pre-formed intermediate to generate α-aceto-α-hydroxyvalerate. Second,
acetohydroxy acid isomeroreductase catalyzes the conversion of α-aceto-α-hydroxyvalerate to α,β-
dihydroxy-β-methylcaproate. Third, dihydroxyacid dehydratase drives dehydratation of the α,β-
dihydroxy acid intermediate to yield α-keto-β-metylcaproate. Fourth, transaminase B catalyzes
transamination of α-keto-β-metylcaproate yo form β-methylnorleucine by using glutamate as donor
of the amino group (Muramatsu et al., 2003).
Homoisoleucine can be also synthesized by the keto acid chain elongation reaction catalyzed by the
leu operon (leuABCD) and the subsequent transamination of α-keto-β-methylvalerate, a metabolic
intermediate of the isoleucine biosynthetic route (Kisumi et al., 1976, I and II).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 27
Figure 5. Schematic representation of the ncBCAA biosynthetic pathway. Gene and enzyme names displayed in
this figure are the same than the ones described in Figure 4. See text for description. Adapted from Reitz
(2017).
2.3.3 Substrate promiscuity of enzymes involved in ncBCAA biosynthesis
The synthesis and accumulation of non-canonical BCAA results from the low specificity of the leu and
ilv-operon-coded enzymes involved in the BCAA biosynthetic pathway for their substrates. This
explains the sequential keto acid chain elongation from pyruvate to α-ketocaproate over α-
ketobutyrate and α-ketovalerate by the actuation of the leu enzymes α-isopropylmalate synthase
(leuA), β-isoprolylmalate dehydrogenase (leuB) and α-isoprolylmalate isomerase (leuCD). Despite
being α-ketoisovalerate the preferred substrate for α-isopropylmalate synthase, this enzyme also
shows certain affinity towards alternative α-keto acids such as pyruvate, α-ketobutyrate and α-
ketovalerate. Table 5 shows kinetic parameters of α-isopropylmalate synthase with various
substrates. Since kinetic information of E. coli-α-IPMS was not found in the literature, kinetic
parameters of α-IPMS from other organisms were considered.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
28 Literature Review
Table 5. Kinetic parameters of α-IPMS from different microorganisms towards various α-ketoacids. The identity
percentage of the α-IPMS of a certain microorganism with respect to E. coli-α-IPMS is shown.
Substrate
Organism
Identity E.
coli
Parameter
α-KIV
Pyruvate
α-KB
α-KV
Ref.
Salmonella
thyphimurium
93 %
Km (mM)
0.06
10
1.1
-
Kohlhaw et al.
(1969)
Alcaligenes
eutrophus
49 %
Km (mM)
0.06
10
1.8
0.4
Wiegel and
Schlegel (1977)
Mycobacterium
tuberculosis
29 %
Km (mM)
0.03
27.3
1.1
0.8
kcat (s-1)
0.81
0.49
0.84
0.29
Hunter and
Parker (2014)
kcat/Km
(mM-1s-1)
27
0.018
0.764
0.363
Serratia
marcescens
24 %
Km (mM)
0.77
3.4
7.7
9
Kisumi et al.
(1976, I)
Moreover, the three consecutive reactions catalyzed by the ilv enzymes acetohydroxy acid synthase
(ilvBN, ilvGM, ilvIH), acetohydroxy acid isomeroreductase (ilvC) and dihydroxyacid dehydratase (ilvD)
can take place in parallel in the biosynthetic pathways of valine, isoleucine and β-methylnorleucine.
As aforementioned, acetohydroxy acid synthase catalyzes a 2-step enzymatic reaction. In the first
reaction step only pyruvate can bind to AHAS, catalyzing decarboxylation to form an active
intermediate. In the second reaction step three alternative substrates (pyruvate, α-ketobutyrate or
α-ketovalerate) can bind to the pre-formed intermediate to generate three alternative acetohydroxy
acids (Gollop et al., 1989). Pyruvate is the preferred substrate for acetohydroxy acid synthase, but it
is also able to transform the other two. Nevertheless, the complexity of acetohydroxy acid synthase
specificity gets even higher considering that there are three different isoenzymes available, each one
showing different substrate preference and catalytic properties (Table 6).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 29
Table 6. Kinetic and affinity parameters of the three acetohydroxyacid synthase isoenzymes (AHAS I, II and III)
towards various substrates. *The factor R1 characterizes the specificity of enzyme acetohydroxyacid synthase
for α-acetohydroxybutyrate formation in relation to α-acetolactate formation. R1 is defined as R1 =
(vAHB/vAL)/([2-ketobutyrate]/[pyruvate]). The factor R2 characterizes the specificity of enzyme acetohydroxyacid
synthase for α-aceto-α-hydroxyvalerate formation in relation to α-acetolactate formation. R2 is defined as R2 =
(vAHV/vAL)/([2-ketovalerate]/[pyruvate]). a: AHAS from E. coli K-12; b: AHAS from S. typhimurium.
Parameter
AHAS I
AHAS II
AHAS III
Reference
Km (pyruvate) (mM)
1.5a
10.5b
6a
Reviewed by Gollop et al.
(1989)
kcat/Km (pyruvate)
(mM-1s-1)
14a
6.6a
0.36a
Reviewed by Vinogradov
et al. (2006)
R1*
2a
65b
40a
Reviewed by Gollop et al.
(1989)
R2*
<0.1a
2.4b
2.3a
Reviewed by Gollop et al.
(1989)
Transaminase B (ilvE) also reports substrate promiscuity since it catalyzes transamination of leucine,
valine, isoleucine, norleucine, norvaline, β-methylnorvaline and homoisoleucine intermediates (Table
7).
Table 7. Kinetic parameters of transaminase B from E. coli towards various substrates. Data obtained from Yu et
al. (2014).
Substrate
Product
Km (mM)
kcat (s-1)
kcat/Km (mM-1s-1)
α-ketoisocaproate
leucine
0.08
24.7
309
α-ketoisovalerate
valine
0.20
10.5
52.5
α,β-keto-β-methylvalerate
isoleucine
0.07
23.0
329
α-ketovalerate
norvaline
0.60
25.9
43.2
α-ketocaproate
norleucine
0.22
23.1
105
2.3.4 Conditions triggering ncBCAA biosynthesis
According to what has been described in the literature so far, three are the main causes of ncBCAA
production in E. coli-based recombinant protein production processes: overflow metabolism and
mixed-acid fermentation, enzymatic feedback regulation and activity of acetohydroxyacid synthase
isoenzyme II.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
30 Literature Review
2.3.4.1 Overflow metabolism and mixed-acid fermentation
The scale-up of recombinant protein production processes from laboratory scale to large-scale
reactors often lead to numerous problems. For instance, for an E. coli-based recombinant protein
production process a 20% reduction of biomass yield and an increase of by-product formation was
reported by Bylund et al. (1998) when scaling up from 3L to 9m3. Riesenberg et al. (1990) not only
showed a biomass reduction but also a lower product yield when scaling up an E. coli process from
30L to 450L. During fermentation, gradient zones of substrate, dissolved oxygen, pH and other
parameters are formed due to inefficient mixing and E. coli cells respond to these environmental
changes by modulating their metabolism (Schweder et al, 1999). The standard procedure to grow E.
coli for recombinant protein production processes at large-scale consists of a fed-batch operation
under aerobic conditions. Addition of concentrated feeding solution is normally carried out from the
top of the reactor while aeration is done by providing oxygen through a vent located at the bottom
of the reactor. Due to the increased mixing times occurring at large-scale, provided glucose and
oxygen are distributed in the reactor in an inverse gradient manner. Hence, reactor regions close to
the glucose inlet are characterized by high glucose concentrations and low dissolved oxygen while
reactor areas close to the oxygen vent show glucose depletion and high dissolved oxygen. Central
regions of the reactor report intermediate values of glucose and oxygen if compared with upper and
lower areas. Moreover, oxygen limitation in the upper part is exacerbated because cells increase
oxygen consumption under glucose excess. E. coli responds to glucose excess and oxygen limitation
by shifting metabolism from oxidative respiration to mixed-acid fermentation, resulting in overflow
metabolism (Enfors et al., 2001). Under these conditions, not only the mixed-acid fermentation
products accumulate, but also pyruvate (Soini et al., 2008, I). Pyruvate excess present intracellularly
increases the metabolic flux going to ncBCAA biosynthesis through the sequential keto acid chain
elongation from pyruvate to α-ketocaproate over α-ketobutyrate and α-ketovalerate by the
actuation of the leu operon-encoded enzymes (Apostol et al., 1997). This hypothesis is supported by
the observations reported by Soini et al. (2011): the combination of oxygen limitation with a constant
glucose supply in a two-compartment STR-PFR scale-down reactor reported a significant impact on
enhancing norvaline biosynthesis due to pyruvate accumulation in a recombinant E. coli cultivation.
Furthermore, Soini et al. (2008, I) originally reported accumulation of pyruvate-based amino acids
such the ncBCAAs norleucine and norvaline as well as alanine and valine in a standard STR fed-batch
E. coli cultivation under glucose excess and induced oxygen limitation upon a stirrer downshift. It was
proposed that a strong accumulation of pyruvate is a prerequisite for α-ketobutyrate formation since
α-IPMS shows lower affinity towards pyruvate than the other alternative substrates: α-
ketoisovalerate and α-ketobutyrate (Table 5). Similarly, α-ketobutyrate has been proposed as a
prerequisite for ncBCAA biosynthesis, since the affinity of α-IPMS for α-ketobutyrate is around 20-
fold lower compared to its preferred substrate, α-ketoisovalerate (Table 5) (Sycheva et al., 2007).
2.3.4.2 Feedback regulation
Another factor strongly influencing ncBCAA biosynthesis is the de-regulation of the leuABCD operon.
It encodes the enzymes catalyzing the chain elongation of various α-keto acids, generating precursors
for leucine and ncBCAAs biosynthesis. Chain elongation of α-ketoisovalerate to α-ketoisocaproate
directs to leucine biosynthesis. Consecutive chain elongation of pyruvate to α-ketocaproate, via α-
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 31
ketobutyrate and α-ketovalerate ends up with ncBCAA biosynthesis. α-ketovalerate and α-
ketocaproate are the precursors of norvaline and norleucine, respectively. Sycheva et al. (2007)
demonstrated that an up-regulation of the leuABCD operon triggers an increase of norvaline and
norleucine accumulation. Furthermore, transcription of the leu operon is negatively regulated by the
presence of leucine, either via attenuation or via Lrp-regulation (Vitreschak et al., 2006; Lin et al.,
1992). Enzymes encoded by the leu operon are also subjected to leucine-mediated feedback
inhibition (Soper et al., 1976). Accordingly, when leucine is in excess the leuABCD operon is
repressed, thus limiting leucine and ncBCAA biosynthesis. However, conditions leading to lower
intracellular levels of leucine cause a de-repression of the leu operon, hence boosting the metabolic
flux going to leucine and ncBCAA biosynthetic pathways. These conditions are especially
accomplished in E. coli-based leucine-rich recombinant protein production processes, where the
metabolic demand for leucine is enhanced (Bogosian et al., 1989; Fenton et al., 1994; Apostol et al.,
1997). This is of special interest since numerous recombinant proteins present a high leucine-content
(Reitz et al., 2018). For instance, recombinant insulin used in this thesis has a leucine content of
14.5% while an average E. coli protein only contains 8.4%. In addition to the de-regulation of the leu
operon, ncBCAA biosynthesis requires accumulation of α-ketobutyrate as the affinity of α-IPMS for α-
ketobutyrate is lower compared to its preferred substrate, α-ketoisovalerate (Table 5) (Sycheva et
al., 2007).
2.3.4.3 Activity of acetohydroxyacid synthetase
Biermann et al. (2013) reported that the degree of norleucine and norvaline biosynthesis strongly
relies on the E. coli strain selected as recombinant protein expression system since strain E. coli
BL21(DE3) reported a significant low accumulation of ncBCAA with respect to strain E. coli K-12. As
opposed to strain E. coli BL21(DE3), genome sequencing and annotation studies revealed that a two-
base insertion event between base pairs 1250 and 1253 is present in the coding sequence of gene
ilvG in E. coli K-12 strains. This insertion causes a shift of the reading frame and, as a consequence, a
stop codon is formed, resulting in a premature termination of ilvG gene expression (Yoon et al., 2012;
Parekh and Hatfield, 1997; Lawther et al. 1981). AHAS II is encoded by ilvG and it is involved in
biosynthesis of isoleucine starting from α-ketobutyrate and, in contrast to AHAS I and AHAS III, AHAS
II is resistant to valine-mediated feedback inhibition. The absence of AHAS II activity in E. coli K-12
strains leads to valine toxicity: AHAS I and AHAS III are inhibited in the presence of valine and, since
AHAS II cannot be properly synthesized, a drastic reduction of leucine and isoleucine biosynthesis
occurs. Under these conditions, growth behavior is jeopardized (Anderson et al. 2001, Biryukova et
al. 2010) and accumulation of α-ketobutyrate is enhanced, thus resulting in higher ncBCAA
production (Soini et al., 2008, I; Sycheva et al., 2007).
2.3.5 Mechanism of ncBCAA mis-incorporation into proteins during translation
Aminoacyl-tRNA synthetases (aaRSs) catalyze the transference of amino acids to their corresponding
cognate tRNA in 2 steps. The first step consists of an ATP-mediated amino acid activation yielding an
aminoacyl-adenylate intermediate. In the second step, the aminoacyl moiety from the activated
intermediate is transferred to the cognate tRNA thus yielding an aminoacylated tRNA, which can
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
32 Literature Review
then be transferred to the ribosome by the elongation factor Tu (EF-Tu) for subsequent protein
translation (Cvetesic et al., 2012).
The fidelity of protein synthesis counts on the aptitude of aaRSs to charge the appropriate canonical
amino acid onto its corresponding tRNA. Such fidelity can be jeopardized by a number of non-
canonical amino acids, particularly ncBCAA, which are structurally similar to their canonical
equivalents (Martinis et al., 1997). Promiscuity of aaRSs can be appreciated in Table 8, where kinetic
parameters of leuRS toward different amino acids are shown (information extracted from Tang and
Tirrell, 2002).
Table 8. Kinetic parameters of leuRS from E. coli towards various amino acids. Data obtained from Tang and
Tirrel (2002).
Substrate
Km (mM)
kcat (s-1)
kcat/Km (mM-1s-1)
leucine
0.018
2.2
122.22
isoleucine
2.568
0.06
0.023
valine
2.356
0.03
0.013
methionine
2.178
0.08
0.037
norvaline
1.155
1.24
1.074
norleucine
2.516
0.22
0.087
For instance, leucyl-tRNA synthetase (leuRS) must distinguish between leucine and the non-canonical
counterpart norvaline, which only differ by a single methyl group (Apostol et al., 1997) (Table 4).
Homoisoleucine can also compete with leucine for leuRS (Van Deventer et al., 2011). The same
happens with methionyl-tRNA synthetase (metRS), which must discriminate between methionine
and norleucine (Kiick et al., 2001), and isoleucyl-tRNA synthetase (ileRS), which must differenciate
between isoleucine and β-methylnorleucine (Muramatsu et al., 2003). In order to prevent mis-
incorporation of non-canonical amino acids, aaRSs have evolved quality control mechanisms which
allow hydrolysis of misactivated amino acids (pre-transfer editing) or misacylated tRNAs (post-
transfer editing) (Figure 6). For leuRS, minincorporation is mainly prevented by post-transfer editing
while for ileRS pre-transfer editing mechanism is preferred (Chen et al., 2011; Cvetesic et al., 2014).
However, the efficiency of the editing mechanism is controlled by kinetic partitioning between
synthetic and editing reaction. For instance, when kinetic partitioning of norvalyl-tRNALeu between
post-transfer editing hydrolysis at the LeuRS editing site and dissociation from leuRS is altered, an
accumulation of norvalyl-tRNALeu might occur, hence increasing mis-incorporation of norvaline at
leucine positions by the translation machinery, since elongation factor Tu (EF-Tu) does not
discriminate against norvaline (Cvetesic et al., 2012). In addition, Apostol et al. (1997) shown that the
ratio of free norvaline with respect to free leucine strongly determines the degree of norvaline mis-
incorporation in the recombinant protein sequence at leucine positions. Hence, under cultivation
conditions triggering biosynthesis of ncBCAA, those accumulate in the cell so that ratio of ncBCAAs
relative to their canonical counterparts increases and this might alter kinetic partitioning at the
editing site of aaRSs, resulting in a reduced editing efficiency and, as a consequence, a major ncBCAA
mis-incorporation. Tang and Tirrell (2002) revealed importance of editing mechanisms in preventing
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 33
ncBCAA mis-incorporation since when editing capacity is attenuated, mis-incorporation of norvaline
is dramatically boosted.
Figure 6. Strategies of leuRS to discriminate among leucine, isoleucine and norvaline. Isoleucine cannot be
processed by leuRS due to its low affinity towards it. However, other amino acids such as norvaline can be
identified by leuRS instead of leucine. Mischarging of amino acids onto tRNA can be prevented by proofreading
at the editing site located close to the active site of aaRS. In most cases mischarged tRNA can be hydrolysed at
editing site of aaRS but, under particular conditions, editing mechanisms work inneficiently, hence triggering
mis-incorporation during translation. Adapted from Cvetesic et al. (2013) and de Pouplana (2014).
2.3.6 Inconvenients of ncBCAA biosynthesis and mis-incorporation into recombinant
proteins
Biosynthesis of ncBCAA can have negative effects in the physiology of E. coli. For instance, norleucine
inhibits DNA replication and methylation reactions, causing DNA damage. In addition, norleucine
restricts methionine biosynthesis by inhibiting homoserine succinyltransferase, the first enzyme of
the methionine biosynthetic pathway (summarized by Bogosian et al., 1989).
Mis-incorporation of ncBCAA during translation can lead to the production of altered proteins,
having non optimal characteristics e.g. altered biological activity, modulated sensitivity to proteolysis
and immunogenicity (Laird and Veeravalli, 2013). There are plenty of evidences showing that the mis-
incorporation of norleucine into recombinant proteins by E. coli may lead to alteration of protein
structure and biological properties. Recombinant norleucine-substituted β-galactosidase was more
resistant to alkylation (Naider et al., 1972). Recombinant norleucine-rich adenylate kinase shows a
much higher resistance to hydrogen peroxide inactivation under denaturing conditions than the
methionine-rich variant (Gilles et al., 1988). Recombinant norleucine-rich mammalian calmodulin
(CaM) presents a decreased enzymatic activity if compared to the normal variant (Yuan and Vogel,
1999). Recombinant norleucine-rich cytochrome P 450 BM-3 heme domain shows an increased
peroxygenase activity than its normal methionine-containing counterpart (Cirino et al., 2003).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
34 Literature Review
When recombinant proteins are to be used as pharmaceutical products for human use, a number of
quality criteria have to be fulfilled in order to ensure its effectiveness and safety when delivered in
the market. Hence, a number of product parameters have to be tested and these should meet the
specifications. One of those parameters is the purity level of the product. The mis-incorporation of
ncBCAA into the product during the recombinant protein production process would generate a pool
of protein variants differing by only a few amino acids. During downstream processing removal of
undesired variants should be performed in order to satisfy the purity specifications. However, this is
sometimes very difficult since most protein variants would show similar properties and sensitivity of
the analytic methods employed might not be enough. This issue represents an important concern for
the pharmaceutical industry since undesired protein variants might have negative effects in patients
(Harris and Kilby, 2014).
2.3.7 Strategies to avoid ncBCAA mis-incorporation into recombinant proteins
Different strategies have been used in order to reduce the degree of non-canonical BCAA mis-
incorporation into recombinant proteins expressed in E. coli, especially for the case of norleucine: (1)
mutating methionine codons of the gene that encodes the recombinant protein so that no
methionine residues are present (Brunner et al., 1997), (2) co-expressing enzymes that are capable of
degrading non-canonical BCAA (Bogosian et al., 2013), (3) supplementing the cultivation medium
with exogenous canonical amino acids to reduce the likelihood of the non-canonical counterpart to
be selected by the corresponding aminoacyl-tRNA synthetase (Apostol et al., 1997; Tsai et al., 1988;
Bogosian et al., 1989; Fenton et al., 1994; Randhawa et al., 1994; Brunner et al., 1997; Abu-Absi et
al., 2008), (4) overproducing methionine by mutating genes involved in methionine biosynthesis and
regulation such as metA, metK and metJ (Usuda and Kurahashi, 2005; Veeravalli et al., 2015; Laird
and Veeravalli, 2013) and in threonine biosynthesis such as thrB and thrC (Usuda and Kurahashi,
2005), (5) knocking-out genes that are involved in the non-canonical BCAA biosynthetic pathway such
as ilvA and leu operon (Tsai et al., 1988; Bogosian et al., 1989; Fenton et al., 1994), (6) optimizing
operational conditions during fermentation (Ni et al., 2015), (7) supplementing trace elements
molybdenum, selenium and nickel in the cultivation medium (Biermann et al., 2013) and (8) using
alternative E. coli expression strains less prone to non-canonical BCAA mis-incorporation (Ni et al.,
2015).
However, most of these strategies present numerous disadvantages. Substitution of methionine
codons for other ones eliminates possibility of norleucine to be mis-incorporated into the
recombinant protein. Nevertheless, alteration of the native amino acid sequence of the target
protein might cause structural changes in the protein conformation leading to reduction of the
protein activity.
Co-expression of ncBCAA-degrading enzymes represents a powerful tool to reduce mis-incorporation
of ncBCAA since no cBCAA supplementation is necessary. It was demonstrated that wild type and
mutant versions of phenylalanine dehydrogenase, valine dehydrogenase, leucine dehydrogenase and
glutamate dehydrogenase from numerous species of bacteria and other organisms are capable of
degrading ncBCAA via NAD+-dependent oxidative deamination (reviewed by Bogosian et al., 2013),
which consists in removing the alpha-amino group from amino acids, generating then α-ketoacids,
ammonia and NADH. However, these enzymes are not only specific for the ncBCAA but are also able
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 35
to degrade the canonical counterparts to a certain extent, which can lead to lower recombinant
protein yields. Hence, in order to make this strategy feasible, more studies concerning protein
engineering should be done in order to increase specificity of the degrading enzymes against the
ncBCAA.
The strategy based on canonical amino acid supplementation works effectively to ensure product
quality since it dramatically reduces the likelihood of ncBCAA to be recognised by the corresponding
aminoacyl-tRNA synthetase. Moreover, addition of leucine not only reduces mis-incorporation of
norvaline due to a higher competitiveness of leucine for leucyl-tRNA synthetase but also a reduction
of norleucine mis-incorporation, since leucine represses the leu operon at both transcriptional and
enzymatic level, hence reducing the metabolic flux going to ncBCAA synthesis. However, this
approach is only feasible at small-scale since in industrial scale production processes additional
medium supplementation would excessively increase process costs and it might also hinder
operation. In addition, media supplementation would dilute fermentation content thus having a
negative effect on cell density and product yield.
Overproduction of methionine by mutated versions of metA, metK and metJ genes overcomes all
aforementioned inconvenients since it was demonstrated that growth behavior and recombinant
protein production yields in some of the mutant strains are close to the control non-modified strain
(Laird and Veeravalli, 2013). This strategy consists in generating new E. coli strains which are
feedback resistant against methionine and S-adenosylmethionine (SAM), which allosterically inhibit
activity of the methionine biosynthetic enzyme encoded by metA (homoserine transsuccinylase). A
substitution of tyrosine by cysteine at position 294 (Y294C) in the amino acid sequence of
homoserine transsuccinylase resulted in a feedback-resistant enzyme (Laird and Veeravalli, 2013).
The enzyme encoded by metK (methionine adenosyltransferase) transforms methionine into SAM.
Both mutation consisting of a substitution of valine by glutamic acid at amino acid position 185
(V185E) and mutation consisting of a deletion of cytosine at base position 1132 in the DNA sequence
of metK resulted in a partially non-functional methionine adenosyltransferase (Laird and Veeravalli,
2013). These new metK variants would result in low levels of SAM and consequently higher de-
repression of the methionine biosynthetic enzymes encoded by metA, metB, metC and metH or
metE. MetJ (encoded by metJ) is a repressor regulatory protein which, when bound to its co-
repressor SAM, represses methionine biosynthetic enzymes at transcriptional level. Both mutation
consisting of a substitution of serine by asparagine at amino acid position 54 (S54N) (Nakamori et al.,
1999) of MetJ and mutation consisting of a complete metJ disruption (Usuda and Kurahashi, 2005)
resulted in de-repression of methionine biosynthetic enzymes and methionine overproduction.
Moreover, Usuda and Kurahashi (2005) reported an increased methionine excretion in E. coli when
using a strain where threonine biosynthetic genes thrB and thrC were disrupted from the genome.
Homoserine kinase (thrB) catalyzes phosphorylation of homoserine to homoserine phosphate while
threonine synthetase (thrC) converts homoserine phosphate to threonine. Since homoserine
transsuccinylase (metA) and homoserine kinase (thrB) compete for homoserine, hence a threonine
auxotrophic strain is expected to trigger methionine overproduction. Engineered E. coli strains
containing one or a combination of mentioned metA, metK, metJ and thrBC mutations would then
result in a significant reduction of norleucine mis-incorporation. Despite the many advantages of this
strategy with respect to previous ones, it is noteworthy to highlight that through methionine
overproduction only norleucine mis-incorporation would be reduced while norvaline and β-
methylnorleucine would still represent a problem for product quality.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
36 Literature Review
L-threonine dehydratase, encoded by gene ilvA, catalyzes the anaerobic formation of α-ketobutyrate
from threonine in a two-step reaction. Since α-ketobutyrate is the initial substrate for ncBCAA
biosynthesis, ilvA inactivation would then result in a reduced ncBCAA production. The leu operon,
encoding for the leucine biosynthetic enzymes (leuA, leuB, leuC and leuD), also plays an important
role in the ncBCAA biosynthesis since it catalyzes a number of enzymatic reactions (the so-called keto
acid chain elongation process) from pyruvate to α-ketovalerate and α-ketocaproate, which are the
precursors of norvaline and norleucine, respectively. Hence, inactivation of one or more genes of the
leu operon would also decrease ncBCAA production. Fenton et al. (1994) demonstrated that by using
E. coli strain CV512, which harbors the disrupting leuA371 mutation, for recombinant protein
production, no norleucine was detected. Bogosian et al. (1989) also reported that E.coli strain LBB24,
harboring a deletion of the entire leu operon, could significantly reduce both norleucine and
norvaline synthesis. However, both described strains required leucine supplementation to grow.
Moreover, Bogosian et al. (1989) generated E. coli strain LBB254, which contains an ilvA disruptive
mutation. Strain LBB254 was not able to grow without isoleucine supplementation. However, after
induction of recombinant protein expression the effect of ilvA mutation was suppressed since strain
was able to grow on minimal medium lacking isoleucine. This is explained because the amino acid
sequence of the expressed recombinant protein (bovine somatotropin) has high leucine content (15
%) if compared with an average E. coli protein (9 %) so that after induction of protein expression
leucine is depleted and, in consequence, transcription of the leu operon is activated, hence
recovering formation of α-ketobutyrate, leucine, isoleucine and ncBCAA. In order to avoid the
suppression effect reported in strain LBB254, Bogosian et al. (1989) generated E. coli strain LBB259,
which lacks both ilvA and leu operon. This strain required leucine and isoleucine for growth. Purified
bovine somatotropin expressed by strain LBB254 showed no norleucine incorporation.
Optimization of operational conditions and nutrient composition during fermentation can also lead
to reduction of ncBCAA production. Ni et al. (2015) demonstrated that by using the Korz-based
chemically defined medium in combination with a DO-stat fed-batch control a significant decrease of
norvaline and norleucine concentrations was achieved without the need to supplement exogenous
canonical amino acids.
Biermann et al., 2013 demonstrated that under oxygen limitation and high glucose concentration
conditions, media supplementation with trace elements molybdenum, selenium and nickel reduces
norvaline and norleucine accumulation. These mentioned elements act as cofactors for the formate
hydrogenlyase complex (FHL). FHL is a membrane-associated metalloprotein complex including a
cytoplasmic molybdenum- and selenium-dependent formate dehydrogenase H (FdhF) and a
transmembrane hydrogenase 3 complex consisting of 6 subunits: the [NiFe] hydrogenase component
HycE, whose active site contains nickel and iron, the iron-sulfur proteins HycB, HycF and HycG and
the integral membrane proteins HycC and HycD. FHL catalyzes formate oxidation to form CO2 and H2
under anaerobic conditions. According to Soini et al. (2008, I) both norvaline and norleucine
accumulate in E. coli cultivations under oxygen limitation and glucose excess. Under those
conditions, glucose catabolism is enhanced and an intracellular accumulation of pyruvate occurs. This
pyruvate excess leads, in turn, to a higher formation of α-ketobutyrate and, in consequence, of
ncBCAA. A scenario triggering even higher ncBCAA formation under aforementioned conditions
happens when inducing expression of leucine-rich recombinant proteins since leucine depletion
would de-regulate leu operon thus increasing further the metabolic flux from pyruvate to ncBCAAs.
FHL, together with lactate dehydrogenase (LDH) and pyruvate formate lyase (PFL), plays a key role in
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 37
the anaerobic metabolism of pyruvate. When using molybdenum, selenium and nickel, functionality
of FHL is enhanced and part of the pyruvate excess generated can then be anaerobically metabolized
to the detriment of the metabolic pathway catalyzed by the leu operon, which ends in ncBCAA
biosynthesis. Both Biermann et al. (2013) and Soini et al. (2008, II) reported a reduction of formate
accumulation in E. coli cultivations after supplementation of the aforementioned trace elements,
which is directly related to a higher FHL functionality.
Moreover, Biermann et al., 2013 reported that the degree of norleucine and norvaline biosynthesis
strongly relies on the E. coli strain selected as recombinant protein expression system since strain E.
coli BL21(DE3) reported a significant low accumulation of ncBCAA with respect to strain E. coli K-12.
As opposed to strain E. coli BL21(DE3), genome sequencing and annotation studies revealed that a
two-base insertion event between base pairs 1250 and 1253 is present in the coding sequence of
gene ilvG in E. coli K-12 strains. This insertion causes a shift of the reading frame and, as a
consequence, a stop codon is formed, resulting in a premature termination of ilvG gene expression
(Yoon et al., 2012; Parekh and Hatfield, 1997; Lawther et al. 1981). AHAS II is involved in biosynthesis
of isoleucine starting from α-ketobutyrate and, in contrast to AHAS I and AHAS III, AHAS II is resistant
to valine-mediated feedback inhibition. The absence of AHAS II activity in E. coli K-12 strains leads to
valine toxicity: AHAS I and AHAS III are inhibited in the presence of valine and, since AHAS II cannot
be properly synthesized, a drastic reduction of leucine and isoleucine biosynthesis occurs. Under
these conditions, growth behavior is jeopardized (Anderson et al. 2001, Biryukova et al. 2010) and
accumulation of α-ketobutyrate is enhanced, thus resulting in higher ncBCAA production (Soini et al.,
2008, I; Sycheva et al., 2007).
Further strategies aiming reduction of ncBCAA mis-incorporation, which are not yet available in the
state of the art, might be considered: (i) substitution of codons reporting higher ncBCAA mis-
incorporation from the coding region of a gene of interest, (ii) protein engineering to increase
specificity of biosynthetic enzymes for certain α-keto acids, (iii) protein engineering to increase
specificity of aminoacyl tRNA synthetases for the canonical amino acids, and (iv) genetic regulation of
target genes involved in the BCAA biosynthetic pathway.
Cvetesic et al. (2016) reported that norvaline is preferentially mis-incorporated at CTG codons, which
represent 47 % of the total leucine codons available in an E. coli cell. They also demonstrated that
leucine codons TTA, TTG and CTT show the most significant reduction of norvaline incorporation.
Two options were hypothesized in order to explain that observations: (i) different aminoacylation
kinetics for each tRNAleu or (ii) different interaction of each Nva-tRNAleu with the translation
machinery. Accordingly, a substitution of CTG codons from the coding region of the gene of interest
to others showing less mistranslation degree might lead to a reduction of norvaline mis-
incorporation in the recombinant protein. Nevertheless, this might also in parallel decrease
recombinant protein productivity since half of the total tRNAleu present in E. coli contain anticodon
CAG (Dong et al., 1996), which might negatively affect translation rate. As for leucine, isoleucine can
be encoded by alternative codons: ATT, ATC and ATA. However, no literature is available showing
preference of β-methylnorleucine mis-incorporation for one of the mentioned codons. Additionally,
Veeravalli et al. (2015) demonstrated that, contrary to previous reports (Bogosian et al., 1989),
norleucine mis-incorporation is not random since, at certain amino acid positions, higher levels of
norleucine are reported. However, this observation cannot be explained by alternative codon usage
since only codon ATG is available for methionine.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
38 Literature Review
α-isopropylmalate synthase (α-IPMS) participates in the so called α-ketoacid chain elongation
pathway. Although the main substrate for α-IPMS is α-ketoisovalerate (α-KIV), the enzyme also
shows certain affinity towards other α-ketoacids such as pyruvate, α-ketobutyrate (α-KB) or α-
ketovalerate (α-KV) (Table 5). This enzymatic promiscuity facilitates biosynthesis of ncBCAA. Similar
to α-IPMS, α-acetohydroxyacid synthase (AHAS) has broad substrate specificities (Table 6).
Depending on the isoenzyme (AHAS I, II or III), AHAS can transform 2 pyruvate molecules into α-
acetolactate and/or pyruvate and α-ketobutyrate into α-acetohydroxybutyrate. However, AHAS can
also convert pyruvate and α-ketovalerate into α-aceto-α-hydroxyvalerate, which is a precursor of β-
methylnorleucine (Muramatsu et al., 2003; Gollop et al. 1989). Accordingly, protein engineering
approaches leading to reduce promiscuity of the leucine and valine/isoleucine biosynthetic enzymes
would be considered as strategy to decrease chain elongation of α-ketoacids leading to biosynthesis
of ncBCAA.
Moreover, as discussed by Veeravalli et al. (2015), usage of aminoacyl tRNA synthetases from other
organisms which cannot use ncBCAA as substrate or engineering of aminoacyl tRNA synthetases to
increase their specificity against cBCAA could be promising approaches to prevent mis-incorporation
of ncBCAA into recombinant proteins manufactured in E. coli. This option might be feasible in the
near future since there is increasing knowledge of the aminoacyl tRNA synthetase structure and its
editing mechanisms.
Another approach which has not yet been explored is the expression regulation of genes involved in
the BCAA biosynthetic pathway in order to minimize ncBCAA biosynthesis and this strategy is the
focus of the current thesis.
2.4 Recombineering in E. coli
Recombineering refers to recombination-mediated genetic engineering. Recombineering comprises
molecular techniques which enable modification of a DNA target by homologous recombination.
Homologous recombination is defined as the pairing of homologous sequences (regions of identical
sequence) between target DNA and linear donor DNA so that the DNA sequence located between
homologous sequences is replaced, resulting in recombinant DNA. Linear DNA that is either single-
(ss-) or double-stranded (-ds) can be used as donor DNA for recombineering. ssDNA are normally
synthetic oligonucleotides while dsDNA can be easily obtained by PCR. A variety of genetic
modifications can be carried out depending on design of the linear donor DNA: gene replacements,
deletions and point mutations. This study focuses on in-frame gene knock-outs (Court et al., 2002).
Homologous recombination naturally occurs in E. coli and it plays an important role in repairing DNA
damage during replication. Homologous recombination is mediated by the orquested action of
proteins RecA, RecBCD and RecF, encoded by genes recA, recBCD and recF, respectively. RecA is
recombinase, RecBCD a helicase/nuclease complex and RecF a DNA replication and repair protein.
RecA binds to single-strand DNA sections forming a RecA-DNA complex able to pair with other
homologous DNA regions. When homology is encountered, RecA triggers a DNA exchange between
both strands. RecBCD plays a major role in initiating recombination at double-strand DNA boundaries
by creating 3’-single-strand overhangs upon degradation of 5’-ends. RecBCD also assists RecA in
binding to ssDNA regions. Similar to RecBCD, RecF starts recombination events at dsDNA terminus.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 39
However, RecF recombination activity is around 100-fold lower than RecBCD (summarized by Court
et al., 2002).
Recombineering demands transformation of linear donor DNA into E. coli hosting efficient
homologous recombination machinery that enables exchange of DNA target with donor DNA. The
bacterial-encoded recombination system has been proven to be inefficient for recombination of
exogenous linear DNA. Moreover, transformability of E. coli is low due to degradation of linear DNA
by intracellular exonucleases. Thus, in order to enhance linear DNA recombination efficiency in E.
coli, alternative recombination systems have been explored. The bacteriophage λ-encoded Red
recombination system has been proven to be at least 50-fold more efficient than previuos systems
(Murphy, 1998). The λ Red recombination system comprises Exo, Beta and Gam proteins, encoded by
genes exo, bet and gam, respectively. While Gam inhibits the bacterial endogenous RecBCD nuclease
activity, Exo and Beta triggers homologous recombination. Hence, linear donor DNA is not
degradated and can recombine with the bacterial genome. Exo has 5’→3’-nuclease activity. It
degrades 5’-ends of the linear donor DNA and creates 3’-ssDNA overhangs. Beta can then bind to the
generated ssDNA overhangs and triggers pairing of donor DNA with the homologous DNA region
present in the E. coli genome. Datsenko and Wanner (2000) developed a simple and highly efficient λ
Red-based recombination method to knock-out single genes from the E. coli chrosomome by using
dsDNA PCR-products containing 36 bp homology arms as linear donor DNA. This methodology was
later employed to generate the so-called KEIO collection, which comprises a palette of single gene
knock-out E. coli strains (Baba et al., 2006).
λ Red recombination system can be expressed from a defective prophage integrated in the E. coli
chromosome or from a helper plasmid. In this thesis, the low-copy helper plasmid pKD46 is
employed. pKD46 carries genes exo, bet and gam under the control of an arabinose-inducible
promoter. Plasmid contains a pSC101-based origin of replication (ori), allowing a low-copy number.
This ori is temperature-sensitive so that plasmid curation can be achieved by cultivating hosting E.
coli at 42 °C. Morevoer, pKD46 contains gene araC, which encodes the repressor AraC protein, thus
preventing basal expression of PBAD promoter and allowing tight expression regulation of λ Red
genes. E. coli K-12 BW25113 was selected in this study for recombineering purposes. Since E. coli K-
12 BW25113 is deficient in arabinose catabolizing enzymes – it contains a deletion of the araBAD
operon: Δ(araD-araB)567 – arabinose can be successfully employed as inducer of λ Red genes
hosted in plasmid pKD46 (Datsenko and Wanner 2010).
In this study, the FRT/FLP homologous recombination system was employed to knock-out operons
ilvIH and ilvBN as well as genes thrA, leuA, ilvC and ilvA from E. coli K-12 BW25113 genome.
2.4.1 Homologous recombination based on FRT/FLP system
The FRT/FLP system originates from Saccharomyces cerevisiae and consists of a site-specific flipase
recombinase (FLP) and the FLP recombinase recognition sites (FRT). FLP triggers site-specific
recombination between FRT sites. This strategy is mainly employed to generate in-frame
chrosomomal gene deletions. In this approach a linear dsDNA fragment containing at both ends
around 36 bp homologous overlaps to the target gene is generated by PCR. Moreover, this dsDNA
fragment carries an antibiotic resistance cassette flanked by two FRT sites. In the current study,
available plasmids pKD3 or pKD4 were used as templates to generate linear donor dsDNA by PCR.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
40 Literature Review
While plasmid pKD3 contains a chloramphenicol-resistance gene flanked by FRT sites and two
priming sites, plasmid pKD4 shares the same features than pKD3, but contains a kanamycin-
resistance gene. PCR primers contain 20 bp homology to the priming sites of the template plasmids
followed by 36 bp ends homologous to the flanking genomic region of the target gene (Datsenko and
Wanner, 2000).
At first, the generated dsDNA fragment is transformed into E. coli and replaces target gene by λ Red-
mediated homologous recombination. λ Red recombination system is normally expressed in a helper
plasmid such as pKD46 (Datsenko and Wanner, 2000). Potential recombinant clones are then
selected for antibiotic resistance. Secondly, antibiotic resistance marker is excised from the genome
by FLP-mediated site-specific recombination between the two FRT sites. FLP is also normally
provided by a helper plasmid such as pCP20. FLP synthesis in plasmid pCP20 is under control of a
temperature-senstive λ repressor and induced at 30 °C (Cherepanov and Wackernagel 1995,
Datsenko and Wanner 2000). Similar to plasmid pKD46, pCP20 contains a temperature-sensitive
pSC101-based origin of replication, allowing easy plasmid curation at 42 °C (Datsenko and Wanner,
2000). Finally, removal of the antibiotic resistance marker leaves in the genome a 103 bp-scar
sequence containing an active FRT site (Figure 7). Accordingly, the FRT genetic scar is considered
suboptimal when several genes need to be altered in the same clone, since with every genetic
modification an additional FRT site would be added to the genome. Since those FRT sites are also
active, the risk of genomic misintegration or even deletions of big chromosal regions increases the
more sequencial modifications are performed. Datsenko and Wanner (2000) recommended using
this strategy for single gene deletions. Reitz (2011) could perform up to 3 knock-outs in a single clone
with this strategy.
Figure 7. Schematic representation of the λ Red-mediated gene replacement with selection marker elimination
by site-specific recombination. See text for description. Adapted from Madyagol et al. (2011).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 41
2.5 Tunable promoter systems in E. coli metabolic engineering
Various tunable promoter systems have been used for high-level expression of recombinant proteins
in E. coli. Nevertheless, an excessive, non-physiological, presence of recombinant protein can lead to
non-desirable effects, such as the generation of inclusion bodies and cell metabolic burden. Tunable
promoter systems are also convenient for metabolic engineering applications where protein
expression at physiological levels is usually required. This section offers a summary of the molecular
mechanisms and applications of four of the most frequently used tunable promoter systems in
metabolic engineering of E. coli (araBAD, rhaBAD, lac and xylS/Pm) (Terpe, 2006), as wells as
inherent advantages and drawbacks.
2.5.1 Introductory considerations
In order to achieve high-level expression of recombinant protein in E. coli, genes encoding the
recombinant protein are usually cloned under the control of inducible promoters on high-copy
plasmids. However, an excessive recombinant protein expression can lead to non-desirable effects,
such as unwanted aggregation and cell metabolic burden. As a result of the forced metabolic burden,
cell physiology might be noticeably affected, being specific growth rate and protein yield the main
altered parameters. In these conditions, plasmid free cells tend to grow faster and plasmid stability is
harmed. Metabolic burden effects are more severe the bigger the plasmid is and the more plasmids
are present in the cell (Glick, 1995; Wu et al., 2016). Therefore, the use of tunable promoter systems
in medium- or low-copy plasmids would reduce the metabolic burden imposed to the cells, thus
minimizing the problems mentioned before (Keasling, 1999; Jones et al., 2000).
Metabolic engineering is the application of genetic engineering in order to refashion the metabolism
of a certain organism to improve cellular properties and to increase the production of a target
metabolite. Metabolic engineering includes the optimization of existing metabolic pathways as well
as the introduction of new ones. The field of metabolic engineering appeared in 1991 (Bailey, 1991).
Since then, plenty of literature is available, demonstrating success of metabolic engineering
approaches in E. coli. The commercialization of the chemical building blocks 1,3-propanediol and 1,4-
butanediol manufactured by engineered E. coli are two examples of success (Nielsen et al., 2016).
Most recent studies achieved the production of L-methionine (Li et al., 2017), monolignol (Chen et
al., 2017), astragalin (Pei et al., 2016), a plant-base anti-inflammatory agent (Ahmadi et al., 2016), L-
malate (Dong et al., 2016) and ectoine (Ning et al., 2016) through metabolic engineering. For
successful metabolic engineering, genetic tools employed should fulfill a number of requirements: (i)
genetic construct carrying the genes should be stable in the host over generations and should not
impose cell metabolic burden, (ii) promoter triggering expression of the target genes should be
tightly regulated, being expression totally repressed in absence of inducer and activated in a linear
dose dependent manner in its presence, and (iii) expression of target genes should be comparable in
all cells of the culture at any time.
Hence, for metabolic engineering applications, where protein expression at physiological levels is
usually required, use of low-copy number plasmids containing target genes is encouraged for proof-
of-concept applications (Jones and Keasling, 1998). Alternatively, integration of the target genes into
the chromosome would improve gene stability and minimize metabolic burden. As opposed to
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
42 Literature Review
plasmid systems, chromosomal integration does not require the use of selective pressure for
maintenance of genetic material. However, expression of genomically integrated genes is
dramatically affected by their position relative to the chromosomal origin of replication (Sousa et al.,
1997; Couturier and Rocha, 2006).
For most metabolic engineering purposes, promoters allowing positive regulation of gene expression
(tunable promoters) in presence of inducer are favored. However, it is important to consider that
inducers are differently transported into the cell and that some of them are more predisposed to be
metabolized. An ideal inducer would be the one entering the cell by passive diffusion and not being
metabolized by the cell machinery, which would allow homogenous and persistent protein
expression in all cells in cultivation, thus avoiding the all-or-none induction effect. As
aforementioned, in the absence of inducer, tunable promoter should be totally repressed while
promoter expression should vary in a positive linear dose dependent manner in its presence. Inducer
should also be inexpensive and non-toxic.
This section offers a summary of the molecular mechanisms and applications of four of the most
frequently used tunable promoter systems in metabolic engineering (araBAD, rhaBAD, lac and
XylS/Pm), as wells as inherent advantages and drawbacks and possible improvements.
2.5.2 araC-PBAD promoter system
The arabinose-inducible araBAD promoter (PBAD) together with its regulator protein AraC has become
a popular tunable expression system for high-level recombinant protein production as well as for
metabolic engineering purposes since expression can be tuned over a broad range of arabinose
concentrations (Guzman et al., 1995).
In the absence of arabinose, AraC coils DNA by binding to araI1 and araO2 thus blocking RNA
polymerase to bind to PC and PBAD promoters and cyclic AMP receptor protein (CRP) to bind to its
respective binding region besides araI1. When arabinose is present in the medium, it is transported
into the cells by the low-affinity high-capacity AraE and the high-affinity low-capacity AraFGH
arabinose transporter systems. If intracellular arabinose concentration surpasses a certain threshold
concentration, it binds to and activates AraC, which in turn undoes the DNA loop, binding to araI2
instead of araO2. This conformation allows binding of CRP and interaction of RNA polymerase with
the promoter regions, thus stimulating expression from promoters PBAD and PC. Promoters PE and PFGH
are also induced when arabinose is present. However, when AraC binds to araO1L and araO1R,
promoter PC is repressed. AraBAD are enzymes involved in arabinose catabolism. Furthermore araC-
PBAD promoter system is also regulated by catabolite repression (Schleif, 2000; Megerle et al., 2008).
Its inherent characteristics make araC-PBAD promoter system especially attractive since basal levels of
expression are particularly low (tight regulation) and response to arabinose is very fast. However,
some literature shows that the expression triggered by this promoter is not homogenous at a cell
population level due to the all-or-none induction mechanism (Siegele und Hu, 1997; Carrier and
Keasling, 1999; Giacalone et al., 2006; Brautaset et al., 2009, Balzer et al., 2013; Binder et al., 2016),
being then problematic when used for metabolic engineering purposes. As mentioned before,
arabinose enters the cell via protein transport systems AraE and AraFGH. However, distribution of
arabinose transporters in a cell population is not homogeneous due to dilution by cell division and
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 43
the arabinose-dependent expression of genes araE and araFGH that causes a stochastic background
expression of promoters PE and PFGH, possibly as a result of a stochastic change in conformation of
AraC while interacting with the promoter region. Consequently, two fractions of cells are present. On
the one hand, the fraction of cells containing enough transporters to accumulate intracellular
arabinose beyond the threshold concentration induces the synthesis of new transporters that uptake
more arabinose, thus amplifying induction. On the other hand, the fraction of cells bearing
insufficient transporters does not accumulate intracellular arabinose beyond the threshold
concentration necessary for induction and hence, cells are not induced. The likelihood that a cell is
induced increases the higher is the arabinose concentration present in the medium (Siegele und Hu,
1997; Morgan-Kiss et al., 2002; Széliová et al., 2016).
Numerous strategies have been proven to be satisfactory for elimination of the all-or-none induction
effect inherent to the araC-PBAD promoter system, allowing then homogenous expression at
population level: (i) deletion of the araFGH operon and substitution of the PE promoter controlling
expression of the AraE transporter by an IPTG inducible promoter (Khlebnikov et al., 2000, 2002), (ii)
deletion of the araBAD and araFGH operons and substitution of the PE promoter controlling
expression of the AraE transporter by a constitutive promoter (Khlebnikov et al., 2001), (iii) deletion
of the araBAD operon and both AraE and AraFGH arabinose transport systems and use of a lactose
transporter (LacY) of relaxed specificity as an L-arabinose transporter (Morgan-Kiss et al., 2002), (iv)
deletion of the araBAD operon and substitution of the PE promoter controlling expression of the AraE
transporter by the stronger PBAD promoter (Széliová et al., 2016) and (v) deletion of the araBAD
operon and use of photocaged arabinose inducers (Binder et al., 2016).
Strategies (i) and (ii) focus on uncoupling expression of the native low affinity-high capacity AraE
transporter from arabinose-dependent regulation. Strategy (iii) completely substitutes the native
arabinose transport systems by an analog transporter, whose expression is uncoupled from
arabinose-dependent regulation. Strategy (iv) allows a general increase of the basal amount of native
arabinose transporters present in the cells, hence decreasing the arabinose concentration threshold
necessary to trigger induction, which in turn results in a more homogenous protein expression at
population level. However, expression of AraE is still dependent on arabinose. Strategy (v) achieves
population homogeneity as photocaged arabinose might enter the cell by passive diffusion, not
depending on transport systems.
2.5.3 rhaBAD promoter system
The rhamnose-inducible rhaBAD promoter together with its regulator proteins RhaR and RhaS has
demonstrated to be tunable over a wide range of rhamnose concentrations (Giacalone et al., 2006).
The rhaBAD promoter shares lots of features with the araBAD promoter. Rhamnose is transported
into the cells by RhaT and it is then metabolized by RhaBAD. When rhamnose is present in the
intracellular medium, it binds to the regulator protein RhaR, which in turn activates transcription of
rhaR and rhaS promoters. RhaR is constitutively expressed at low levels. The regulator protein RhaS
also binds to rhamnose, hence activating transcription from rhaT and rhaBAD operon. Transcription
of rhaBAD promoter is also regulated by catabolite repression, since it requires the binding of CRP
(Egan and Schleif, 1993; Via et al., 1996).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
44 Literature Review
There are also evidences that L-rhamnose promoter exhibits the ‘all-or-none’ response in E. coli
(Ozbudak et al., 2004; Giacalone et al., 2006; Brautaset et al., 2009; Afroz et al., 2014; Kelly et al.,
2016) and other microorganisms (Cardona und Valvano, 2005). L-rhamnose concentration affects the
relative fraction of cells in the fully induced or uninduced states (Afroz et al., 2014; Ozbudak et al.,
2004). Similarly to L-arabinose, L-rhamnose catabolism involves the inducible expression of the
transport system RhaT by rhamnose, so this might explain how rhaBAD promoter induction could
develop bimodality by this mechanism (Kelly et al., 2016). Despite that, no literature is available for
rhaBAD promoter concerning possible strategies minimizing the all-or-none induction effect but a
feasible option would be the use of strains defective in rhamnose-catabolizing enzymes.
In order to successfully use the rhaBAD promoter for metabolic engineering purposes, it might be
convenient to use genetically engineered E. coli strains where the rhaBAD operon is depleted, hence
avoiding metabolization of rhamnose. In addition, similarly to the araBAD promoter, the reported all-
or-none induction mechanism may be solved by uncoupling expression of the native RhaT
transporter from rhamnose-dependent regulation; i.e. substitution of the native promoter
controlling expression of the RhaT transporter by a constitutive promoter.
2.5.4 lac promoter system
The lactose operon is formed by genes lacZ, lacY and lacA. Lactose is transported into the cells by a
lactose permease, encoded by lacY. Lactose is then metabolized into glucose and galactose by β-
galactosidase, which is encoded by lacZ. The gene lacA encodes a galactoside O-acetyltransferase
which might have an important role in cellular detoxification by acetylating non-metabolizable
pyranosides, thus reducing rate of retransport into the cell (Roderick, 2005; Marbach and
Bettenbrock, 2012). lacA does not seem to be involved in lactose metabolism. The lactose operon is
regulated by the LacI repressor, which is encoded by the gene lacI. LacI binds to the operator region
of the operon, thus inhibiting transcription of genes lacZ, lacY and lacA. However, when lactose is
present in the medium, allolactose can bind to the LacI repressor and triggers the release of the
repressor from the operator region, thus activating transcription of genes lacZ, lacY and lacA.
However, binding of the cAMP-bound catabolite activator protein (CAP) is also required for
transcription activation, being then lac operon subjected to catabolite repression (Busby and Ebright,
2001; Görke and Stülke, 2008).
Allolactose is the natural inducer of the lac operon. However, it can be metabolized by β-
galactosidase and this is not advantageous for metabolic engineering applications since a constant
induction of expression of target genes is preferable. However, artificial structural analogs of
allolactose such as TMG (methyl-1-thio-β-d-galactopyranoside) and IPTG (isopropyl β-d-1-
thiogalactopyranoside), that cannot be metabolized, have been developed (Daber et al., 2007;
Marbach and Bettenbrock, 2012). Similar to lactose, both artificial inducers are also transported into
the cell by LacY at low inducer concentrations. Nevertheless, at high inducer concentrations an
alternative transport independent on LacY is possible as well. As aforementioned, LacA can acetylate
TMG and IPTG, thus becoming inactive as inducer molecules (Marbach and Bettenbrock, 2012).
Numerous variants of the original E. coli lac promoter have been designed for molecular biology
applications: lacUV5, Ptac, Ptrc and T7lac. The lacUV5 promoter contains 2 point mutations in the -10
promoter region and an additional one at -66, thus differing in only 1 nucleotide from the consensus
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Literature Review 45
sequence (Hirschel et al., 1980). As a result, the lacUV5 promoter features a higher affinity for E. coli
RNA polymerase than the wild type lac promoter. The lac operon driven by the UV5 promoter
achieves about 50 fold higher expression than the wild type version (Pribnow, 1975) and does not
require the presence of cAMP-CAP, that is, is insensitive to catabolite repression (Arditti et al., 1973;
Silverstone et al., 1970). Ptac is a hybrid promoter where -35 region is derived from the trp promoter
while -10 region comes from the lac UV5 promoter, being about 10 times more efficient than the
parental lac UV5 promoter (de Boer et al., 1983). Ptrc is a variant of Ptac where one additional base
pair was inserted between -35 and -10 promoter regions, having about 90% the activity of Ptac. The
T7lac promoter is a variant of the UV5 promoter where the gene encoding T7 RNA polymerase is
under the control of the lacUV5 promoter genomically integrated while the gene of interest is
located in a plasmid under the control of a T7 promoter. This system offers a very tight regulation
and achieves about 4 fold higher expression than the Ptrc promoter (Tegel et al., 2011).
lac-based promoters are specially leaky, that means, expression is also triggered in a basal level even
when no inducer is present. This is because LacI repressor does not tightly bind to the operator and
can dissociate. Hence, when using lac-based promoters in cloning vectors is important to take into
consideration that enough LacI repressor is present in order to avoid leaky expression of the
promoter. The number of LacI molecules necessary to repress lac promoter mainly depends on the
number of operators regulating the promoter, the affinity of the repressor to the operator and the
copy number of the plasmid bearing the lac promoter (Penumetcha et al., 2010). Several strategies
have been developed in order to improve the regulation tightness of lac-based promoters: use of low
copy plasmids for recombinant protein expression, overexpression of LacI repressor by using
stronger lacI promoter variants and use of LacI variants with higher affinity towards operator regions.
Different variants of lacI promoter sequences leading to different expression levels of lacI have been
reported in E. coli strains: lacI+ and lacIq in Calos (1978), I-UJ177 in Calos and Miller (1980) and lacIq1
in Calos and Miller (1981). According to Glascock et al. (1998), an E.coli wild type cell (lacI+ variant)
has about 10 LacI molecules. The lacIq promoter version shows 10-fold expression enhancement of
lac repressor if compared with the lacI+ variant (100 lacI molecules) while expression by lacIq1 is 17-
fold stronger than that of lacIq and 170-fold stronger than that of lacI+ (1700 lacI molecules). In
addition, numerous lacI gene mutants conferring a higher binding affinity of the repressor for the
operator region have been utilized: LacI-I12, LacI-X86 and the double mutant LacI-I12/X86. LacI-I12
and LacI-X86 confers a 50-100 fold improvement of binding affinity if compared with the wild type
LacI while the LacI-I12/X86 variant causes a 10000 fold increase (Penumetcha et al., 2010).
Similarly to the araBAD promoter, the all-or-none induction effect was first observed for the lac
operon by Novick and Weiner (1957). LacY catalyses the uptake of IPTG and TMG, which in turn
induces further expression of LacY, thus leading to a positive feedback loop that is responsible of the
bimodal behavior of gene expression at population level (Ozbudak et al., 2004). However, Marbach
and Bettenbrock (2012) and Binder et al. (2014) demonstrated that a lacY deletion results in a
homogeneous response to induction at population level. However, when using this mutant, a 20-fold
higher TMG and 10-fold higher IPTG concentration was necessary to achieve similar induction levels
than the wild type strain (Marbach and Bettenbrock, 2012).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
46 Literature Review
2.5.5 Pm/XylS promoter system
It has recently become a popular promoter system. Plenty of literature is available proving the
efficiency of the Pm/XylS expression system (reviewed by Brautaset et al., 2009).
XylS is constitutively transcribed by its own promoter Ps2, and the gene of interest is placed under
the control of Pm promoter. The inducer molecule (benzoic acid derivatives) enter the cells via
passive diffusion and bind to XylS, which then activates transcription of the gene of interest from Pm
promoter (Brautaset et al., 2009).
Unlike the previously mentioned lactose, arabinose and rhamnose-based expression systems,
benzoate inducers for the activation of Pm/XylS systems don’t rely on active transport systems but
enter the cells via passive diffusion (Brautaset et al., 2009; Binder et al., 2016). Hence, homogeneous
expression at the cell population level is reported since no regulated transported system is present
(Balzer et al., 2013; Binder et al., 2016) and no previous strain genetic engineering is needed.
Moreover this system is flexible, since it does not require specific features of the host, in contrast to
most other systems. Hence, the performance of this expression system is not strain dependent in
E.coli. Induction is carried out with benzoic acids, which are very cheap and non-toxic. Conventional
induction with m-toluate especially leads to a minimum growth impairment and homogenous
population (Binder et al., 2016). However, the system leads to slight protein expression also in the
absence of inducer and that expression level is especially high for the Pm variant ML1-17 (Balzer et
al., 2013). According to Binder et al. (2016), benzoate induction systems worked better at 30°C.
Pm/XylS expression system is also commercially available (Website 1). This expression system is
considered optimal for metabolic engineering when used in conjunction with minimal replicons such
as RK2. One of these vectors, pJB658, has proven useful for tightly regulated recombinant gene
expression in several gram-negative species (Blatny et al., 1997; Brautaset et al., 2000; Winther-
Larsen et al., 2000 and EP2078076 B1; Sletta et al., 2004).
2.5.6 Summary
To sum up, araBAD, rhaBAD and lac-based promoters are not optimal since they exhibit an ‘all-or-
none’ expression response at cell population level. However, for araBAD and lac promoters,
strategies minimizing that negative effect are available. Hence, in order to successfully implement
these promoters for metabolic engineering strategies, previous strain engineering is required. As
opposed to aforementioned promoters, Pm/XylS promoter shows homogeneous expression
response at population level thanks to diffusion-based transport into the cell so that no previous
strain engineering is necessary. Moreover, Pm/XylS promoter seems promising since expression level
is tunable and the inducer is non-toxic and inexpensive.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Research Hypotheses and Aim of the Project
47
3. Research Hypotheses and Aim of the Project
The incorporation of non-canonical branched chain amino acids (ncBCAA) such as norleucine,
norvaline and β-methylnorleucine into recombinant proteins during E.coli production processes has
become a crucial matter of contention in the pharmaceutical industry, since such mis-incorporation
can lead to the production of altered proteins, having non optimal characteristics . Although several
strategies have been developed in order to reduce ncBCAA mis-incorporation in E. coli, a number of
limitations impede them to be effectively applied in large-scale recombinant protein production
processes. This thesis focuses on another approach which has not yet been explored and that might
contribute to close the scientific gap.
The main objective of this thesis is the generation of new genetically engineered E. coli strains
allowing the production of recombinant proteins with a considerably reduced content of ncBCAA.
Since ncBCAA biosynthesis is controlled by enzymes encoded by genes involved in the BCAA
biosynthetic pathway such as the leu- and ilv-encoded enzymes, it would be valuable to have a more
comprehensive understanding of the effect of the expression regulation of the implicated genes into
ncBCAA biosynthesis. Hence, the novel E. coli strains would enable exogenous tunable expression of
target genes involved in the BCAA biosynthetic pathway so that the expression level of a certain
target gene reporting a minimized ncBCAA biosynthesis and mis-incorporation could be identified.
Screening of the engineered E. coli strains was performed in a mini-reactor platform by cultivating
them in fed-batch mode under standard cultivation conditions. Expression levels of each target gene
were determined by the amount of L-arabinose added in each experimental case. At the end of the
cultivation, ncBCAA concentrations were determined in the inclusion body and intracellular soluble
fraction so that for each tested target gene the expression level resulting in the further reduction of
ncBCAA production could be identified.
However, since potential novel E. coli strains are aimed to be used in recombinant protein
production processes taking place in large-scale reactors, it might be that data resulting from
screening in mini-reactors under aforementioned cultivation conditions would not be conclusive
since large-scale effects are not properly reproduced in that experimental design. As cultivation
screenings cannot be directly performed at pilot or production scale due to capacity and economical
reasons, application in the mini-reactor system of scale-down approaches simulating environmental
perturbations occurring at large-scale would be suitable to obtain more realistic data. One of the
main causes triggering ncBCAA biosynthesis in E. coli is the metabolic shift from oxidative respiration
to overflow metabolism driven by inefficient mixing in large-scale reactors (Enfors et al., 2001).
Under these conditions, not only the mixed-acid fermentation products accumulate, but also
pyruvate (Soini et al., 2008, I). Pyruvate excess present intracellularly increases the metabolic flux
going to ncBCAA biosynthesis through the sequential keto acid chain elongation from pyruvate to α-
ketocaproate over α-ketobutyrate and α-ketovalerate by the actuation of the leu operon-encoded
enzymes (Apostol et al., 1997). Hence, another objective of this thesis was to apply in the mini-
reactor system cultivation conditions triggering ncBCAA biosynthesis, as it happens in industrial scale.
A bis dato non-explored cultivation strategy combining pyruvate pulses and oxygen limitation was
employed in this work in order to reproduce large-scale effects. This novel scale-down approach may
have an advantage with respect to other strategies recently described in single compartment
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
48
Research Hypotheses and Aim of the Project
reactors consisting on glucose pulses combined with oxygen limitation (Anane et al., 2019) or oxygen
down-shift combined with glucose excess (Soini et al., 2008, I).
After screening of the engineered E. coli strains in the mini-reactor system under both standard
cultivation conditions and conditions reproducing large-scale effects i. e. pyruvate pulsing and
oxygen limitation, the most promising E.coli mutants were selected. The last objective of this thesis
was to test those potential E. coli mutants expressing the optimal amount of the target gene in a 15L
reactor under cultivation conditions subjected to pyruvate pulsing and oxygen limitation in order to
verify its superior performance concerning recombinant protein impurity profile in comparison with
the non-engineered E. coli strain.
Considering the aforementioned, the following research hypotheses have been contemplated to
build up the basis of this current investigation:
• Hypothesis 1. Down-regulation of single genes leuA, thrA or ilvA, up-regulation of single
operons ilvBN, ilvIH or ilvGM as well as up-regulation of single gene ilvC would reduce ncBCAA
biosynthesis and subsequent mis-incorporation into recombinant proteins expressed in E. coli.
Taking into consideration the BCAA metabolic pathway, following strategies were pursued by
modulating expression of target genes in E. coli in order to reduce ncBCAA biosynthesis: (i)
limit conversion of pyruvate to α-ketobutyrate, (ii) limit transformation of threonine to α-
ketobutyrate and (iii) limit conversion of α-ketobutyrate to α-ketovalerate. Strategy (i) might
be achieved by down-regulating operon leuABCD but also by up-regulating operon ilvBN,
strategy (ii) may be realised by down-regulating the thr genes as well as ilvA. Strategy (iii)
could be accomplished by down-regulating operon leuABCD and up-regulating ilvIH, ilvGM or
ilvC. According to this hypothesis novel E. coli strain mutants were genetically engineered so
that the expression of single target genes (leuA, thrA, ilvA, ilvC, ilvIH, ilvBN and ilvGM) could
be modulated in order to evaluate the effect of genetic modulation in ncBCAA biosynthesis.
The hypothesis was tested for each target gene by screening each engineered E. coli mutant
expressing different levels of the target gene in a mini-reactor system. During the screening,
the influence of a certain expression level of each target gene in ncBCAA biosynthesis and
mis-incorporation into the recombinant protein was evaluated. Performance of the most
promising E. coli strain mutants selected after screening was additionally verified in a 15L-
reactor scale.
• Hypothesis 2. Cultivation conditions subjected to pyruvate pulses and oxygen limitation
represent a novel strategy reproducing large-scale effects, i.e. biosynthesis of ncBCAA and
overflow metabolism by-products
According to this scenario, pyruvate would enter the cell and directly accumulate there,
hence triggering formate, acetate and ncBCAA biosynthesis.
The hypothesis was verified by comparing levels of formate, acetate and ncBCAA over
cultivation time between the E. coli cultivation performed under standard conditions and the
cultivation subjected to pyruvate pulses and oxygen limitation.
• Hypothesis 3. Cultivation conditions subjected to pyruvate pulses and oxygen limitation
represent a novel strategy reproducing large-scale effects, i.e. biosynthesis of ncBCAA and
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Research Hypotheses and Aim of the Project
49
overflow metabolism by-products, faster than in scale-down cultivation strategies based on
glucose excess.
When using glucose pulses, large-scale effects (i. e. increase of ncBCAA biosynthesis) can first
be reported once glucose is converted to pyruvate and this starts accumulating in the cell.
However, when using pyruvate pulses, pyruvate would directly enter the cell and would
accumulate faster intracellulary, resulting in a rapid trigger of ncBCAA biosynthesis. This
scale-down approach would then accelerate the velocity of the strain screening process.
The hypothesis was verified by comparing the time passed from the moment the
perturbation was applied until a significant increase of ncBCAA concentrations was reported
between the cultivation strategy used in this study (pyruvate pulses and oxygen limitation)
and other published approaches based on glucose excess.
• Hypothesis 4. Metabolic effects triggered by cultivation conditions subjected to pyruvate
pulses and oxygen limitation translate first in the cytosol as ncBCAA biosynthesis and
accumulation and then in the recombinant protein as ncBCAA mis-incorporation.
NcBCAA are first produced by the BCAA biosynthetic pathway in the cytosol and accumulate
intracellularly. After a certain accumulation, ncBCAA can then be mis-incorporated into the
expressed recombinant protein by the translation machinery.
The hypothesis was tested by comparing the time passed from the moment the perturbation
was applied until a significant increase of ncBCAA concentrations was reported in both
inclusion body and intracellular soluble protein fraction.
• Hypothesis 5. The probability that a certain cAA is exchanged by the respective ncBCAA
analog depends on the number of individual positions of such cAA in the sequence of the
expressed recombinant protein.
The amount of leucine residues in the recombinant mini-proinsulin is higher compared to
methionine or isoleucine. Hence, recombinant protein would report higher mis-incorporation
of norvaline, compared to norleucine and β-methylnorleucine.
The hypothesis was verified by comparing the concentrations of each ncBCAA that were
measured in the inclusion body fraction.
• Hypothesis 6. Induction of recombinant mini-proinsulin expression has a positive effect on
ncBCAA biosynthesis and mis-incoporation.
Overexpression of leucine-rich recombinant proteins cause depletion of the intracellular
leucine pool which, in turn, causes de-regulation of the enzymes encoded by the leu operon,
resulting in ncBCAA biosynthesis.
The hypothesis was tested by comparing ncBCAA levels before and after IPTG-mediated
induction of recombinant mini-proinsulin.
• Hypothesis 7. Effective use of lac-based promoters in expression plasmids strongly depends on
the levels of expressed LacI repressor and the E. coli strain employed as expression host.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
50
Research Hypotheses and Aim of the Project
The Ptac promoter was used in this study for expression of recombinant mini-proinsuiln from
plasmid pSW3. The Ptac promoter is inducible by IPTG but, under certain conditions, promoter
is leaky, i.e. expression is also triggered in a basal level even when no inducer is present. This
mainly occurs when there are not sufficient LacI molecules to bind to the operators and
repress basal expression.
The hypothesis was tested by comparing promoter efficiency of Ptac, i. e. levels of leaky
expression and achieved recombinant protein production, cloned in pSW3 plasmid variants
expressing different LacI repressor levels hosted in two different E. coli K-12 strains.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
51
4. Materials & Methods
4.1 Materials, reagents and equipment
The different reagents, materials and equipment used in the current study are described in Table S1,
Table S2 and Table S3, respectively.
4.2 Software
Software employed in this work is listed in Table S4.
4.3 Bacterial strains
Bacterial strains employed in this study are listed in Table S5.
4.4 Plasmids
An overview of plasmids used in this study and their genetic features is shown in Table S6. Plasmid
maps were generated by Snapgene® software.
4.5 Media
Following media were used in this study:
• Luria-Bertani (LB) Medium (Roth, Cat.Nr: X964.1)
• Luria-Bertani (LB) Agar (Roth, Cat.Nr: X965.1)
• Super Optimal broth with Catabolite repression (SOC) (NEB, Cat.Nr.: B9020S)
• M9 medium
M9 medium contained (per 400 mL): 0.8 mL 1 M MgSO4, 0.4 mL 0.1 M CaCl2, 4 mL 40 % glucose, 8 mL
10 mg/mL thiamin, 16 mL 10 % casamino acids and 370 mL salt solution. The salt solution comprised
(per L): 7.5 g Na2HPO4. 2H2O, 3 g KH2PO4, 0.5 g NaCl and 1 g NH4Cl.
• Modified Davis and Mingioli medium
Composition of modified Davis and Mingioli medium was as follows: 7 g/L K2HPO4, 3 g/L KH2PO4, 0.5
g/L Na citrate.2H2O, 5 g/L (NH4)2SO4, 4 g/L glucose, 0.1 g/L MgSO4.7H2O and 100 μL/L 10000 X trace
elements solution. The 10000 X trace elements solution comprised (per 10 mL): 5 mL 0.1 M
FeCl3.6H2O, 0.2 mL 1 M CaCl2.2H2O, 0.1 mL 1 M MnCl2.4H2O, 0.1 mL 1 M ZnSO4.7H2O, 0.1 mL 0.2 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
52
Materials & Methods
CoCl2.6H2O, 0.2 mL 0.1 M CuCl2.2H2O, 0.1 mL 0.2 M NiCl2.6H2O, 0.2 mL 0.1 M Na2MoO4.5H2O and 0.2
mL 0.1 M H3BO3.
• TUB medium
Composition of TUB mineral salt medium was as follows: 2 g/L Na2SO4, 2.468 g/L (NH4)2SO4, 0.5 g/L
NH4Cl, 14.6 g/L K2HPO4, 3.6 g/L NaH2PO4.2H2O and 1 g/L (NH4)2-H-citrate. The mineral salt medium
was then supplemented with 2 mL/L MgSO4 solution (1.0 M) and 2 mL/L trace elements solution. The
trace element solution comprised (per L): 0.5 g CaCl2 x 2H2O, 0.18 g ZnSO4 x 7H2O, 0.1 g MnSO4 x H2O,
16.7g FeCl3 x 6H2O, 0.16 g CuSO4 x 5H2O, 0.18 g CoCl2 x 6H2O.
4.6 Primers
Snapgene® software was employed in designing primers for both PCR and DNA sequencing purposes.
Chemically synthesized primers were provided by ThermoFisher Scientific. An overview of the
primers used in this study is summarized in Table S7.
4.7 Standard molecular biology methods
4.7.1 PCR
Two different DNA polymerases were employed, depending on the PCR objective. If PCR was
destinated to generate DNA fragments for subsequent cloning or sequencing reactions, the Phusion
High-Fidelity DNA Polymerase was employed, since it has proofreading ability. PCR mix and
thermocycling conditions when using Phusion High-Fidelity DNA Polymerase are shown in Table 9
and Table 10, respectively.
Table 9. PCR mix for Phusion® High-Fidelity DNA Polymerase.
PCR Component
50 μL reaction
Final concentration
Nuclease-free water
To 50 μL
-
5X Phusion HF or GC Buffer
10 μL
1X
10 mM dNTPs
1 μL
200 μM
10 μM Forward primer
2.5 μL
0.5 μM
10 μM Reverse primer
2.5 μL
0.5 μM
Template DNA
*Var.
-
DMSO (optional)
(1.5 μL)
3%
Phusion DNA Polymerase
0.5 μL
1 U/50 μL PCR
Total volume
50 μL
-
*Var.: Recommended amounts of DNA template for a 50 μL reaction are as follows:
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
53
DNA
Amount
Genomic
50 ng - 250 ng
Plasmid or viral
1 pg - 10 ng
Table 10. Thermocycling conditions for Phusion-based PCR. Ta: annealing temperature; ∞: infinite.
Step
Temperature (°C)
Time (min:s)
Number of cycles
Initial denaturation
98
0:30
1
Denaturation
98
0:10
5
Ta1
variable
0:30
Extension
72
0:15 - 0:30 /kb
Denaturation
98
0:10
30
Ta2
variable
0:30
Extension
72
0:15 - 0:30 /kb
Final extension
72
10:00
1
Hold
4
∞
As opposed to standard primers, overlapping primers are primers containing a homologous region to
template DNA as well as a non-homologous region at 5’. When using overlapping primers for PCR,
two different annealing temperatures were employed. Ta1 was applied during the first 5 PCR cycles
and was calculated taking into account only the homologous primer regions. Ta2 was applied in the
last 30 PCR cycles and was calculated considering the whole primer sequence. If Ta1 and Ta2 were
equal, same temperature was used for all cycles (routine PCR). Extt was calculated based on fragment
size and nature of template. However, when PCR was performed for molecular clone verification
purposes, where the exact sequence of the PCR product is not crutial, the so-called colony PCR
variant was used in combination with the faster DreamTaq polymerase. In this case, a colony was
resuspended in 20 μL of nuclease-free water and the mix was vortexed and heated at 95°C for 5
minutes to generate a cell lysate. Then, suspension was centrifuged for 2 min at 5000 g and 1 µL of
supernatant was used as PCR template. PCR mix and thermocycling conditions when using DreamTaq
DNA Polymerase are shown in Table 11 and Table 12, respectively. Nevertheless, some PCRs done in
this study combined usage of both mentioned polymerases. In this case, thermocycling conditions of
the DreamTaq polymerase were applied. PCRs were carried out with thermocycler SimpliAmp™
Thermal Cycler (Thermo Scientific). Calculation of annealing temperatures was performed by using
software SnapGene (GSL Biotech LLC) and Tm Calculator (NEB).
Table 11. PCR mix for DreamTaq DNA Polymerase.
PCR component
50 μL reaction
Final concentration
Nuclease-free water
to 50 μL
-
10X DreamTaq Buffer
5 μL
1X
10 mM dNTPs
1 μL
200 μM
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
54
Materials & Methods
10 μM Forward primer
2.5 μL
0.5 μM
10 μM Reverse primer
2.5 μL
0.5 μM
Template DNA
*Var.
-
DreamTaq DNA Polymerase
0.625 μL
1.25 U/50 μL PCR
Total volume
50 μL
-
*Var.: Recommended amounts of DNA template for a 50 μL reaction are as follows:
DNA
Amount
Genomic
100 ng – 1 μg
Plasmid or viral
10 pg – 1 ng
Table 12. Thermocycling conditions for colony PCR using DreamTaq polymerase. Tm: melting temperature; ∞:
infinite.
Step
Temperature (°C)
Time (min)
Number of cycles
Initial denaturation
95
3:00
1
Denaturation
95
0:30
40
Annealing
Tm– 5
0:30
Extension
72
variable
Final extension
72
15:00
1
Hold
4
∞
4.7.2 Plasmid extraction (mini and midiprep)
A 5 mL LB medium containing the corresponding antibiotics was inoculated with a single colony of a
certain E. coli strain. Culture was then incubated overnight at 30 or 37 °C and 220 rpm, overnight.
Next day, plasmid purification was performed according to the protocol provided in Qiaprep spin
miniprep kit (Qiagen, Ref: 27106). After plasmid isolation, DNA concentration and purity was
measured by Nanodrop. However, for low copy plasmids, such as the pACG_araBAD and
pACG_XylSPm plasmid variants, the QIAGEN Plasmid Plus Midi Kit (Qiagen; Ref.: ID12943) was
employed. In that case, 50 mL LB containing the corresponding antibiotics was inoculated with a
single colony of a certain E. coli strain. Culture was then incubated at 30 or 37 °C and 220 rpm,
overnight. Next day, plasmid purification was performed by using the QIAGEN Plasmid Plus Midi Kit
(Qiagen; Ref.: ID12943), according to the provided protocol. After plasmid isolation, DNA
concentration and purity was determined by Nanodrop.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
55
4.7.3 Genomic DNA extraction
Genomic DNA was extracted according to the following protocol, adapted from Packeiser et al.
(2013). Cells from a colony were picked up from agar plates by using a sterile pipette tip and
resuspended into 20 μl of 10 mM Tris / 1 mM EDTA (TE) buffer. To facilitate cell disruption, the
mixture was vortexed for 10 s and incubated at 98 °C for 5 min. The lysate was next microcentrifuged
(1 min, 10000 g) and the resulting supernatant was collected, diluted with distilled water at a
1:5∼1:20 dilution ratio, and subjected to downstream applications.
4.7.4 Preparation of E. coli electro-competent cells
Depending on the exogenous DNA type to be transformed (linear dsDNA or plasmid DNA) and the
features of the E. coli host (presence or not of plasmid pKD46), two different adapted protocols were
employed in this study.
4.7.4.1 General protocol for plasmid electroporation into E. coli cells
20 mL of sterile LB medium were introduced into a 100 mL Erlenmeyer flask and medium was
inoculated with a single colony of the corresponding E. coli host. Cells were incubated at 30-37 °C and
220 rpm, overnight. Next day, 100 mL of sterile LB medium were filled into a 500 mL Erlenmeyer flask
and medium was inoculated with 1 mL of the fresh overnight culture. When the culture reached an
OD600nm of 0.4-0.5, cells were harvested as two 50 mL aliquots in sterile Falcon tubes and those were
left on ice for 15 min. Tubes were then centrifuged for 15 min at 2500 g and 4 °C. After
centrifugation, as much supernatant as possible was discarded and cell pellets were resuspended in
25 mL ice-cold dH2O. Tubes were filled up to 50 mL with ice-cold dH2O and cells were left on ice again
for 15 min. Tubes were centrifuged again as described before. As much supernatant as possible was
removed, pellet was resuspended in 25 mL ice-cold dH2O, tubes were filled up to 50 mL and cells
were incubated on ice for 15 minutes. Tubes were centrifuged again as described before. Now after
discarding the supernatant, cells were resuspended in a total volume of 25 mL ice-cold dH2O, both
tubes were pooled into one tube, tube was filled up to 50 mL and cells were left on ice for 10 min. A
last centrifugation step was performed as described before. Cells were resuspended in 1 mL ice-cold
10 % glycerol and 50 μL aliquots were prepared in sterile 1.5 mL Eppendorf tubes. Tubes were
maintained in ice until electroporation took place or they were stored at -80 °C. Protocoll was
adapted from Reitz (2011).
4.7.4.2 Protocol for linear dsDNA transformation into E. coli cells containing plasmid pKD46
20 mL of sterile LB medium supplemented with 100 µg/mL ampicillin were introduced into a 100 mL
Erlenmeyer flask and medium was inoculated with a single colony of E. coli BW25113 containing
pKD46. Cells were incubated at 30 °C and 220 rpm, overnight. Next day, 100 mL of sterile LB medium
supplemented with 100 µg/mL ampicillin were filled into a 500 mL Erlenmeyer flask and medium was
inoculated with 1 mL of the fresh overnight culture. When the culture reached an OD600nm of 0.1,
expression of the phage red recombinase genes present in pKD46 were induced by adding 8 mL of a
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
56
Materials & Methods
1.5 M L-arabinose solution to reach a final arabinose concentration of 100 mM. Cells were then
incubated again until the culture reached an OD600nm of 0.7. Cells were harvested as two 50 mL
aliquots in sterile Falcon tubes and those were left on ice for 15 min. Tubes were then centrifuged for
15 min at 2500 g and 4 °C. After centrifugation, as much supernatant as possible was discarded and
cell pellets were resuspended in 25 mL ice-cold dH2O. Tubes were filled up to 50 mL with ice-cold
dH2O and cells were left on ice again for 15 min. Tubes were centrifuged again as described before.
As much supernatant as possible was removed, pellet was resuspended in 25 mL ice-cold dH2O, tubes
were filled up to 50 mL and cells were incubated on ice for 15 minutes. Tubes were centrifuged again
as described before. Now after discarding the supernatant, cells were resuspended in a total volume
of 25 mL ice-cold dH2O, tubes were pooled into one tube, tube was filled up to 50 mL and cells were
left on ice for 10 min. A last centrifugation step was performed as described before. Cells were
resuspended in 1 mL ice-cold 10 % glycerol and 50 μL aliquots were prepared in sterile 1.5 mL
Eppendorf tubes. Tubes were maintained in ice until electroporation took place or they were stored
at -80 °C. Protocoll was adapted from Reitz (2011).
4.7.5 Electroporation of DNA into E. coli electro-competent cells
A 1.5 mL Eppendorf tube containing 950 μl sterile SOC was prepared (for E. coli hosts containing
pKD46, SOC medium was supplemented with 100 mM L-arabinose) and pre-warmed at 30-37 °C. A 2
mm electroporation cuvette was placed on ice and electroporator Gene Pulser Xcell™ (BioRad) was
adjusted to the following conditions: 2500 V, 25 μF, 200 Ω. Electro-competent cells were thawed on
ice (about 10 min) and 1-5 μL of the purified DNA were pipetted into an aliquot of competent cells.
The resulting mix was homogeinized by stirring with the pipette. The mix was transferred into an
electroporation cuvette avoiding introduction of bubbles and making sure that cells deposit across
the bottom of the cuvette. An electrical pulse was then applied. After transformation, the time
constant was checked, which should be between 4.5 and 6.0. The pre-warmed SOC medium was
pipetted into the cuvette immediately and the mix was homogeneized gently with the pipette. The
mix was then transfered to a 15 mL round-bottom tube. Cells were left incubating for 1 h at 37 °C (or
1.5-2 h at 30 °C) and 250 rpm, so that the antibiotic resistance gene can be expressed. 100 μL of the
cell suspension (and/or serial dilutions) were streaked out on LB-plates containing the corresponding
antibiotic for selection. Depending on the plasmid used for transformation, plates were incubated at
30 or 37 °C, overnight. Plates were stored at 4° C. Protocoll was adapted from Reitz (2011).
4.7.6 Curation of temperature-sensitive plasmids
Plasmids pKD46 and pCP20 are characterized by a temperature-sensitive origin of replication. In
order to remove those plasmids from the cell, clones of interest containing one of those plasmids
were streaked out for isolation of single colonies on a LB agar plate and it was incubated at 42 °C for
18 to 20 h. Then numerous singles colonies were picked up from LB-plates and resuspended in 30 μL
dH2O. From the resulting suspension, a 0.5 µL droplet was pipetted in following selection plates: an
LB plate containing ampicillin, an LB plate containing a second selective antibiotic, if required, and an
LB plate without any antibiotic. Plates were incubated overnight at 37 °C. Full ampicillin sensitivity
(no colony growth at all) indicated loss of plasmid pKD46 or pCP20. When no or partial ampicillin
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
57
sensitivy was shown (colony grew with difficulty), an additional temperature treatment at 42 °C was
carried out, since pasmid rests could still remain in the cell.
4.7.7 Electrophoresis analytical gel
In this study, gel electrophoresis was used to verify amplification of target fragments after PCR and to
check for appropriate restriction digestion. Therefore, 10 µL of sample (PCR or restriction products)
were mixed with 2 μL Gel Loading Dye Purple (6X) (NEB, Ref: B7024). 10 μL of the resulting mix were
then loaded on a E-Gel® 1.2 % Agarose (EtBr stained, Thermo Scientific, Cat.Nr.:G501801). 5 μL of the
molecular marker Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb) (NEB, Ref: N0469) were used. DNA
electrophoresis was run for 30 minutes at 100 V. Gel was then visualized using Gel Doc™ EZ Imager
(BioRad).
4.7.8 PCR spin purification
QIAquick PCR Purification Kit (Qiagen; Ref.: 28104) was employed according to the provided protocol
in order to purify PCR and restriction reaction products. This procedure performed better than gel
purification since DNA loss during purification was much lower. However, this purification strategy
was only employed when PCR product was really pure in the analytic gel, in order to avoid
contamination with other non-desired DNA molecules.
4.7.9 Extraction from DNA gels
When the analytical gel revealed an unspecific PCR reaction (a smear of bands corresponding to PCR
products of different sizes), purification of the target band corresponding to the correct PCR product
was performed by extraction from a preparative gel. Hence, a 100 mL TAE buffer-based preparative
gel containig 1% agarose and 0.5 µL 10 mg/mL ethidium bromide was prepared, DNA sample was
loaded and gel was run at 180 V for 40 min. The target band was then cut and purified from the gel
by using the QIAquick Gel Extraction Kit (Qiagen, Ref: 28706), according to the provided protocol.
4.7.10 Standard cloning based on restriction digestion and ligation
Restriction enzymes cleave DNA at or close to specific recognition sites. In this study, restriction
enzymes were used for 3 different purposes: plasmid cloning, restriction analysis of potential clones
and removal of PCR templates from PCR products by DpnI. DpnI is a restriction enzyme cutting
methylated DNA. This enzyme is interesting when transforming linear PCR products into E. coli. DpnI
treatment of PCR products ensures elimination of template plasmid DNA, thus reducing the
likelihood of false positives on transformation plates. Restriction enzymes and its corresponding
buffer were acquired from New England Biolabs (NEB). The setup for restriction digestion is shown in
Table 13. The HF (high-fidelity)–variant of a restriction enzyme allows faster restriction (15 minutes
at 37 °C) than the standard one (1 h at 37 °C).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
58
Materials & Methods
Table 13. Setup of a standard restriction reaction. Incubation temperature, time and buffer is enzyme-
dependent. RE: restriction enzyme.
Component
Amount for 50 μL reaction
DNA
1 μg
RE1-HF
0,5 μL
RE2-HF
0,5 μL
10X CutSmart Buffer
5 μL
Nuclease-free water
Up to 50 μL
After restriction of vector and insert with restriction enzymes for cloning purposes, T4 DNA Ligase
was used to join both fragments. Ligation reaction setup was according to Table 14. Ligation was
performed with the Quick Ligation Kit (NEB, Ref: M2200), according to the protocol provided. The
standard 3:1 insert:vector molar ratio was employed in this study. Ligation mix was incubated at 25°C
for 5 minutes, left on ice and 2 μL of the ligation product were transformed into 50 μL NEB 5-alpha
competent E.coli (high efficiency) cells, according to the protocol provided (NEB, Ref: C2987H).
Table 14. Setup of a standard ligation reaction.
Component
Amount for 20 μL reaction
Quick ligase reaction
buffer 2X
10 μL
Vector
100 ng
Insert
*
Nuclease-free water
Up to 20 μL
Quick ligase
1 μL
*Website 2 was used to calculate the optimal volume of insert necessary for ligation considering vector and
insert sizes and the insert:vector ratio.
4.7.11 InFusion cloning
InFusion cloning is an advanced cloning approach allowing generation of complex DNA constructs by
assemblying multiple single fragments. The main advantage of fhis cloning strategy is that it does not
require the usage of restriction enzymes or ligase. Single DNA fragments are joined and cloned into a
linearized vector in a single and fast step. Each DNA fragment participating in the InFusion cloning
reaction contains a 15 bp overlap homologous to the linearized vector or to the adjacent DNA
fragment. Enzymes involved in the InFusion cloning reactions generate 15 bp ssDNA overhangs at the
5’ end of both vector and DNA linear fragments which are then annealed at the complementary sites.
This novel cloning strategy was employed by using the In-Fusion® HD Cloning Kit (Takara; Ref.:
638909), according to the protocol provided. The InFusion cloning reaction was set up as shown in
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
59
Table 15. The InFusion cloning mix was incubated for 15 min at 50 °C and the resulting InFusion
cloning product was transformed into competent E. coli NEB5α cells (NEB; Ref.: C2987I), according to
provided procotol.
Table 15. Setup of the InFusion cloning mix.
Compound
Volume (µL)
5X InFusion HD Enzyme Premix
2
Linearized vector
*
Fragment 1
*
Fragment 2
*
Nuclease-free water
up to 10 µL
*The on-line tool In-Fusion® Molar Ratio Calculator (Website 3) was used to calculate optimal amounts
(volumes) of vector and insert for the In-Fusion® Cloning reaction considering fragment size and a relation
vector:inserts of 1:2.
4.7.12 Mutagenesis
The kit “Phusion site-directed mutagenesis kit” (Thermoscientific, Ref: F-541) allows to introduce
point mutations as well as deletions and insertions to available plasmids during PCR. In this strategy
the entire original plasmid is amplified using 5’-phosphorylated primers which introduce the desired
mutations at the target sites. The amplified linear PCR product, which now contains the desired
mutation, is then autoligated in a 5-minute reaction.
The mix setup and thermocycling conditions employed in this study for the mutagenesis PCR reaction
were as aforementioned in Table 9 and Table 10, respectively. Ligation was performed with the Quick
Ligation Kit (NEB, Ref: M2200) by incubating ligation mix at 25 °C for 5 minutes and ligation product
was transformed into 50 μL NEB 5-alpha competent E.coli (high efficiency) cells according to the
protocol provided (NEB, Ref: C2987H).
4.7.13 Sequencing
The target DNA fragment to be sequenced can be present in a plasmid or in a PCR product. When
target DNA was contained in a plasmid, sequencing mix was prepared as indicated in Table 16.
Table 16. Setup of sequencing reaction using plasmid DNA as template.
Component
Volume (µL)
Primer 10 μM
1
Plasmid DNA (300 ng)
Variable
Nuclease-free water
Adjust to 6
When target DNA was present in a PCR product, 5 μL of such PCR product were mixed with 2 μL of
Exo-Star Reagent (Sigma, Ref: GEUS78210) and the resulting mix was incubated for 15 minutes at 37
°C and then for 15 minutes at 80 °C. From the resulting mix 2 μL were used for the sequencing
reaction as described in Table 17. The ExoStar treatment, which comprises enzymes exonuclease I
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
60
Materials & Methods
and alkaline phosphatase, removes unbound primers and dNTPs from PCR products. Betaine
improves sequencing performance of difficult DNA templates containing GC-rich base pairs, guanine
stretches or GCT-type repeats (Haqqi et al., 2002). Sequencing was performed internally at Sanofi-
Aventis Deutschland GmbH.
Table 17. Setup of sequencing reaction using a PCR product as template.
Component
Volume (µL)
Primer 10 μM
1 μL
Exo-Star-treated PCR product
2 μL
Betaine 5M
2 μL
Nuclease-free water
1 μL
4.7.14 Preparation of bacterial glycerol stocks
A molecularly verified positive isolated colony of the corresponding strain was picked from a fresh LB
plate and resuspended in a 4 mL LB liquid culture containing the corresponding antibiotic and it was
grown at 30-37 °C at 220 rpm, overnight. 1 mL of sterile 50 % glycerol was added to 1 mL of the
overnight culture in labelled 2 mL cryogenic vials. Cryovials were then stored at -80°C.
4.7.15 SDS-PAGE
Following components were mixed in order to prepare a 200 µL 1X sample buffer solution (per
sample): 20 μL of 10X Bolt/NuPAGE Sample reducing agent, 50 μL Bolt/NuPAGE LDS sample buffer 4X
and 130 μL dH2O. OD1/mL cell pellets were resuspended with 200 µL 1X sample buffer to give a
concentration of OD5/200 μL. Cell disruption was performed by sonication using VialTweeter block
sonotrode attached to ultrasonic processor UP200St (standard settings: capacity: 100%, amplitude:
0.9 sec, duration: 1 min). Incubation was done for 10 minutes at 95°C. Afterwards, samples were
shortly spinned and chilled down. 10-20 μL were loaded per lane (0.25-0.5 OD/lane) into a
Bolt/NuPAGE Bis-Tris Puls gel containing 12% polyacrylamide. 10 μL of SeeBlue® Plus2 Pre-stained
Protein Standard was also loaded. 1X Bolt/NuPAGE MES SDS Running Buffer was used as running
buffer. 500 μL of Bolt/NuPAGE Antioxidant were loaded into the inner gel chamber. Electrophoresis
was carried out at 190 V during 45 min. After electrophoresis, gel was left in a plastic tray and
washed with dH2O. Protein staining was then performed with Instant Blue (20 mL) for about 60 min,
in agitation. Gel was again washed and de-stained for 3 times with dH2O (each time 5-10 min, in
agitation). Gel was then visualized using Gel Doc™ EZ Imager (BioRad).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
61
4.8 Amino acid analysis by GC-FID
4.8.1 Intracellular soluble protein fraction and inclusion body isolation from cell extracts
The BugBuster Protein Extraction Reagent (Novagen, Ref: TB245) was used for isolation of
Intracellular soluble protein and inclusion body fractions from cell extracts according to the following
protocol:
1. OD600nm of the culture was determined. A certain volume from each culture was taken and
cells were harvested from liquid culture by centrifugation at 10000 g for 10 min. Supernatant
was then decanted and removed from the pellet.
2. Cell pellet was resuspended in room temperature-BugBuster reagent by pipetting or gentle
vortexing, using 5 mL reagent per gram of wet cell paste.
3. To reduce viscosity of the lysate and improve protein extraction, 10 µL Lysonase were added
per gram wet cell paste. The cell suspension was mixed gently and it was incubated at room
temperature on a shaking platform or rotating mixer at a slow setting for 20 minutes.
4. Insoluble cell debris was removed by centrifugation at 16000 g for 20 min at 4 °C. The
supernatant was transfered to a fresh tube and it was maintained on ice for short term
storage (2-3 h) of frozen at -20°C until needed. Supernatant corresponded to the intracellular
soluble protein fraction, including free amino acids. Pellet was saved for further inclusion
body purification.
5. Pellet was resuspended in the same volume of BugBuster Protein Extraction Reagent that
was used to resuspend the original cell paste (step 2). Suspension was then pipetted up and
down or vortexed. A complete resuspension of the cell pellet solubilizes and removes
contamination proteins and is critical to obtain a high purity preparation.
6. 10 µL Lysonase were added per gram wet cell paste, suspension was mixed gently by
vortexing and it was incubated at room temperature for 5 minutes.
7. An equal volume of 1:10 diluted BugBuster reagent (in deionized water) was added to the
suspension and this was vortexed for 1 min.
8. The suspension was centrifuged at 5000g for 15 min at 4°C in order to collect the inclusion
bodies. Supernatant was removed with a pipette.
9. The inclusion bodies were resuspended in 1:10 diluted BugBuster (10 mL per gram of original
cell paste), vortexed and centrifuged as described in step 8.
10. Step 9 was repeated twice for a total of 3 washes with 1:10 diluted BugBuster.
11. Pellet was resuspended once more in 1:10 diluted BugBuster (10 mL per gram of original cell
paste).
12. In order to check purification level of inclusion bodies containing the recombinant protein, an
SDS-PAGE was carried out with the inclusion body samples. A certain volume corresponding
to an OD1/mL was taken from the solution containing purified inclusion bodies obtained at
step 11 and sample was mixed with 50 µL of Bolt LDS Sample Buffer (4X) and 20 µL of Bolt
Reducing Agent (10X) and it was filled up to 200 µL with deionized water (OD5/200µL).
Samples were heated at 70°C for 10 minutes and 20 µL were loaded into a Bolt Bis-Tris Puls
gel containing 12% polyacrylamide gel (0.5 OD/lane). 10 μL of SeeBlue® Plus2 Pre-stained
Protein Standard was also loaded. 1X Bolt MES SDS Running Buffer was employed as running
buffer. 500 μL of Bolt Antioxidant were loaded into the inner gel chamber. Electrophoresis
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
62
Materials & Methods
was carried out at 190 V during 40 min. After electrophoresis, gel was left in a plastic tray
and washed with dH2O. Protein staining was then performed with Instant Blue (20 mL) for
about 60 min, in agitation. Gel was again washed and de-stained for 3 times with dH2O (each
time 5-10 min, in agitation). Gel was then visualized using Gel Doc™ EZ Imager (BioRad).
13. The remaining IB suspension was centrifuged at 16000 g for 15 min at 4 °C and supernatant
was discarded. Purified IB pellets were then stored at -80 °C.
4.8.2 Hydrolysis of intracellular soluble protein fraction and inclusion bodies
A given sample volume (max. 250 µL) was mixed up to 1 mL with 5M HCl. Resulting solutions were
introduced in crystal vials with screw caps which are resistant to aggressive acids like HCl. Vials were
incubated closed for 24h at 80 °C for acid hydrolysis. Afterwards, vials were left opened in a heating
block for 16-24h at 65 °C while rotating until all liquid was evaporated. Hydrolyzed samples were
then resuspended with dH2O or a solution containing 20 mM HCl and 10% isopropanol and resulting
solutions were used for further amino acid isolation.
4.8.3 ncBCAA analysis with EZ:faastTM for free (physiological) amino acid analysis by GC-
FID
Amino acid isolation from samples was performed according to the following protocol, provided by
the “EZ:faastTM for free (physiological) amino acid analysis by GC-FID” kit:
1. Preparation of Elution Medium: 3 volume parts of Reagent 3A were combined with 2 volume
parts of Reagent 3B. For example, for 12 samples, 1.5 mL Reagent 3A and 1 mL Reagent 3B
were mixed.
2. For calibration purposes, a certain volume of calibration standard solutions (SD1, SD2 and
ncBCAA solutions) were mixed with 100 μL 200 µM L-2-aminobutyric acid (ABA). For analysis
of samples, a certain volume of the hydrolyzed protein sample (inclusion body or intracellular
protein solube fraction) was mixed with 100 μL 200 µM L-2-aminobutyric acid (ABA). In order
to identify retention times of canonical and non-canonical amino acids, a solution containing
following components was additionally prepared for each GC-run set: 25 µL of 200 µM
standard SD1, 25 µL of 200 µM standard SD2, 25 µL of 200 µM norvaline, 25 µL of 200 µM
norleucine and 25 µL of 200 µM β-methylnorleucine. Standard solutions SD1 and SD2 were
provided in the kit and contained canonical amino acids. ABA was used as internal standard
amino acid in order to perform calculations.
3. Liquid samples generated at step 2 were introduced into provided sample preparation vials.
A sorbent tip was then attached to a 1.5 mL syringe and piston was pulled back slowly (within
1 min) to pass total sample solution from the vial through sorbent tip.
4. 200 µL Reagent 2 (Washing Solution) were pipetted into the sample preparation vial.
5. Washing Solution was passed slowly through sorbent tip and into syringe barrel with the 1.5
mL syringe as well.
6. Sorbent tip with the bound amino acids was dettached from the syringe and left it in the vial.
Liquid accumulated in the syringe was discarded and syringe was washed with a 1:1
isopropanol:dH2O solution for another use.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
63
7. 200 µL Elution Medium from step 1 were pipetted into the sample preparation vial.
8. A 0.6 mL syringe was taken and piston was pulled back halfway up the barrel. Sorbent tip
with the bound amino acids was then attached to that syringe.
9. Elution Medium was passed slowly through the sorbent tip until liquid reached the white
filter plug located in the tip.
10. Liquid and sorbent particles were then ejected out of the tip into the vial by pushing the
piston down fast.
11. 50 µL Reagent 4 (derivatization reagent) were pipetted with the provided Microdispenser
into the vial.
12. The liquid was emulsified in the vial by vortexing for about 10-15 s and vial was left for 1 min
at least in order to complete derivatization reaction. This step was repeated once.
13. 100 µL Reagent 5 were pipetted into sample vial with the provided Microdispenser, solution
was vortexed for 10-15 s and reaction was left to proceed for an additional minute.
14. 100 µL Reagent 6 were pipetted into sample vial (without Microdispenser) and solution was
vortexed for about 5 s. A centrifugation step at 3000 rpm for 1 min was performed to
improve separation. At the end of the process, the emulsion should have separated into two
layers. The upper layer contained the isolated amino acids.
4.8.4 ncBCAA analysis by gas chromatography-flame ionization detection (GC-FID)
After amino acid isolation process, around 120 μL of the resulting upper layer were introduced into
GC vials and 2 μL were then injected into the GC analyzer. The GC was run according to following
oven conditions: equilibration time of 0.5 min, 110 °C for 1 min, 30 °C/min heating up to 320 °C and
then 320 °C for 1 min. Each GC run was about 9 min. Nitrogen was used as a carrier gas with a
constant flow rate of 1.5 mL/min. Injection was carried out with a 1:15 split ratio at 250 °C.
4.8.5 Data analysis of data generated by GC-FID
For data analysis the software “Agilent Chemstation, Rev. B.03.02 [341]” was employed. For each GC
run, a chromatogram containing peaks corresponding to different amino acids was generated. Most
of the peaks were automatically detected and integrated by the software. However, some other
smaller or more complex peaks, including the ones corresponding to the ncBCAA, had to be manually
selected and integrated. After integration, a table was generated, where information and features
about individual selected peaks (e.g. retention time, peak height, peak area,…) was summarized. ABA
was selected as the internal standard for amino acid data analysis so that all individual amino acid
calculations were normalized to ABA. Thus, peak area of each individual peak corresponding to one
amino acid was divided by the peak area of ABA, obtaining a ratio. Those generated ratios were then
interpolated into previously elaborated calibration curves and the equivalent amino acid
concentration was estimated. From those concentrations, and knowing the volume of the samples,
amino acid mass was then calculated. Those amino acid mass values could then be normalized to the
corresponding OD600nm of the cultivation as well as to the corresponding mass of expressed
recombinant protein. In addition, ratio of normalized concentrations of canonical and the
corresponding non-canonical BCAA could also be calculated.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
64
Materials & Methods
4.9 Generation of pSW3 plasmid variants with different LacI expression
levels: pSW3, pSW3_lacI+ and pSW3_lacIq
A set of pSW3 plasmid variants were engineered, where different lacI promoter variants (lacI+ and
lacIq) were cloned together with gene lacI in order to express different amounts of LacI repressor
(sections 4.9.1-4.9.3). The generated pSW3 plasmid variants were then transformed into E.coli K-12
BW25113 for further analysis of mini-proinsulin expression (4.9.4-4.9.6).
4.9.1 Sequencing of pSW3
The complete sequence of plasmid pSW3 was not available. However, after a large research through
different patents and papers (Amann et al., 1983; Habermann, 1986; Dörschung et al., 1989;
Habermann und Wengermayer, 1996), the sequence of some genetic regions present in plasmid
pSW3 was identified. Using those sequences as templates, primers were accordingly generated in
order to complete sequencing of the whole plasmid pSW3. Following primers were designed:
pSW3_F1_seq, pSW3_R1_seq, pSW3_R2_seq, pSW3_F2_seq, pSW3_F3_seq, pSW3_R3_seq,
pSW3_F4_seq, pSW3_R4_seq, pSW3_F5_seq, pSW3_R5_seq, pSW3_F6_seq, pSW3_R6_seq,
pSW3_F7_seq, pSW3_R7_seq and pSW3_F8_seq. After sequencing, resulting sequencing reads were
aligned (Figure S50) and the sequence and genetic map of whole pSW3 could be generated thanks to
the multiple overlapping reads (Figure S2A). It was confirmed that pSW3 does not contain the rop
gene. Hence, plasmid pSW3 is high-copy number (115) but not medium-copy number (39-55).
Plasmid pSW3 lacks the bom site, which prevents transmission by conjugation, and contains the
enhanced lacIq promoter and an incomplete, non-functional lacI gene.
4.9.2 Cloning lacIq into pSW3 to generate plasmid pSW3_lacIq
The sequence of lacI+ and lacIq promoter variants were obtained from Glascock and Weickert (1998)
and the sequence of lacI was obtained from pGEX-4T-1 (Figure S4A). Cloning of lacIq promoter variant
and gene lacI into pSW3 was carried out through restriction with enzymes PaeI and PciI. First of all,
fragment containing lacIq promoter variant plus gene lacI, flanked by PaeI and PciI restriction sites,
was synthetically generated by GeneArt, resulting in the plasmid
16ADFSUP_2034902_lacIq_promotor (Figure S3A). Afterwards, restriction with PaeI and PciI of both
pSW3 (vector) and 16ADFSUP_2034902_lacIq_promotor (insert) was carried out and resulting
fragments were ligated, thus resulting plasmid pSW3_lacIq (Figure 8, Figure S2C). Ligation mix was
then transformed into NEB 5-alpha competent E.coli (high efficiency) cells.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
65
Figure 8. Cloning procedure followed to generate pSW3_lacIq. Generated with Snapgene®.
20 colonies from the transformation plates were verified by colony PCR using primers pSW3_R3_seq
and pSW3_R6_seq. PCR products were tested by gel electrophoresis (Figure S7), being all tested
clones positive with the exception of clones 4, 9, 10, 12 and 13, which shown an additional non-
expected band. In order to ensure that sequence of plasmids from positive clones was correct, some
reactions were prepared for sequencing by using primers pSW3_R3_seq and pSW3_R6_seq and
colony PCR product as template. pSW3_lacIq sequence was correct for both tested clones 3 and 5 E.
coli NEB5α pSW3_lacIq (Figure S51). A bacterial glycerol stock was prepared for clone 3 E. coli NEB5α
pSW3_lacIq.
4.9.3 Mutation of pSW3_lacIq to generate pSW3_lacI+
Plasmid pSW3_lacI+ (Figure S2B) was generated by introducing a point mutation into pSW3_lacIq
through PCR according to section 4.7.12. Following 5’-phosphorylated primers were used for
mutagenesis PCR: Mut_lacI+_F and Mut_lacI+_R. Mutagenesis PCR was tested under different buffer
combinations (buffer HF alone, buffer HF plus DMSO, buffer GC alone and buffer GC plus DMSO).
Mutagenesis-PCR products were tested by gel electrophoresis (Figure S8). For the 4 tested PCR
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
66
Materials & Methods
conditions, the expected band was obtained. For the subsequent ligation step, PCR product obtained
by using the HF buffer was employed. Ligation mix was then transformed into NEB 5-alpha
competent E.coli (high efficiency) cells. 9 colonies from the transformation plates were verified by
colony PCR using primers pSW3_R3_seq and pSW3_R6_seq. PCR products were tested by gel
electrophoresis (Figure S9), being all tested clones positive. In order to certify if mutation in the lacI
promoter region of the 9 tested clones occurred successfully, a reaction mix was prepared for
sequencing. Primer pSW3_R3_seq was used as sequencing primer while colony PCR product was
used as template. pSW3_lacI+ sequence was correct for clone 1 E. coli NEB5α pSW3_lacI+ (Figure
S52). A bacterial glycerol stock was prepared for clone 1 E. coli NEB5α pSW3_lacI+.
4.9.4 Transformation of pSW3 into E. coli K-12 BW25113 and molecular verification
Electro-competent E. coli K-12 BW25113 cells were prepared and transformed with plasmid pSW3 by
electroporation. In order to molecularly confirm transformation of plasmid pSW3, various clones
growing in transformation plates were selected and a miniprep was done in order to purify plasmids.
Verification of pSW3 presence was performed by restriction analysis using EcoRI. Digestion products
were tested by gel electrophoresis (Figure S10), confirming that all 6 tested clones contained plasmid
pSW3. A bacterial glycerol stock was prepared for clone 1 E. coli K-12 BW25113 pSW3.
4.9.5 Transformation of pSW3_lacIq into E. coli K-12 BW25113 and molecular verification
Electro-competent E. coli K-12 BW25113 cells were prepared and transformed with plasmid
pSW3_lacIq by electroporation. Some colonies from the transformation plates were verified by
colony PCR using primers pSW3_R3_seq and pSW3_R6_seq. PCR products were tested by gel
electrophoresis (Figure S11), revealing that all 8 tested clones contained plasmid pSW3_lacIq. A
bacterial glycerol stock was prepared for clone 1 E. coli K-12 BW25113 pSW3_lacIq.
4.9.6 Transformation of pSW3_lacI+ into E. coli K-12 BW25113 and molecular verification
Electro-competent E. coli K-12 BW25113 cells were prepared and transformed with plasmid
pSW3_lacI+ by electroporation. Some colonies from the transformation plates were verified by
colony PCR using primers pSW3_R3_seq and pSW3_R6_seq. PCR products were tested by gel
electrophoresis (Figure S12), confirming that all 5 tested clones contained plasmid pSW3_lacI+. A
bacterial glycerol stock was prepared for clone 1 E. coli K-12 BW25113 pSW3_lacI+.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
67
4.10 Design of an arabinose-based tunable expression plasmid
(pACG_araBAD)
A tunable expression plasmid under the control of the araBAD promoter system (pACG_araBAD) was
engineered by Infusion cloning (sections 4.10.1-4.10.3) and genes leuA, ilvC, ilvA, ilvIH, ilvBN, ilvGM
and thrA were subsequently cloned (section 4.10.4). Resulting plasmids would allow expression
regulation of target genes by exougenous addition of L-arabinose.
4.10.1 Features of the DNA fragments involved in InFusion cloning
The plasmid pACG_araBAD was formed by the junction of 3 different DNA segments. Fragment 1
contained araBAD promoter, gene araC, multicloning site (MCS), His-tag sequence and T7-based
terminator. Gene transcription was controlled by a tunable arabinose promoter (Guzman et al.,
1995) and a strong T7 terminator*. AraC is necessary for the activation of the arabinose promoter.
The MCS allowed cloning with rare-cutting restriction enzymes NheI and NotI. Protein was expressed
as a C-terminal 6xhis-tag fusion protein to allow detection by western blot. A gly-gly-gly-gly-ser linker
was also available, connecting protein of interest with 6xhis-tag (Chen et al., 2013). An efficient
ribosome binding site from bacteriophage T7 gene 10 was also present in fragment 1 (Olins and
Rangwala, 1989). Right after C-terminal 6xhis-tag three stop codons were present in order to ensure
proper translation termination. Fragment 2 contained cmR cassette, which confers resistance against
chloramphenicol. Fragment 3 contained origin of replication ori2 and other genes involved in
replication process. Those elements work coordinately, ensuring 1 copy number plasmid per cell,
thus allowing cloning and stable maintenance of very large DNA fragments. Origin of replication was
based on the Ori2 (OriS) replicon of the F (fertility) factor of E. coli, a vector which encodes the SopAB
functions for active partitioning. These functions act at SopC to ensure that each daughter
cellreceives a copy of the plasmid. Initiation factor RepE (also known as RepA) mediates assembly of
a replication complex at Ori2 (Imber et al., 1983; Komori et al., 1999; Mori et al., 1986).
4.10.2 Generation of PCR fragments for InFusion cloning
The first step was to obtain by PCR the 3 abovementioned DNA segments either from plasmids
already available or by chemical synthesis. Fragment 1 was chemically synthesized and subsequently
cloned in plasmid 16ABZ5NP_1934177 (Figure S3B) by GeneArt. Original sequence of araBAD
promoter as well as araC gene was obtained from plasmid pBAD-DEST49 (Figure S4C). Fragment 1
was then amplified by PCR from plasmid 16ABZ5NP_1934177 with primers araBAD_InFusion_F and
araBAD_InFusion_R. Fragment 2 was directly amplified from plasmid pCP20 (Figure S1D) with
primers cmR_InFusion_F and cmR_InFusion_R. Fragment 3 was directly amplified from pETcoco1
(Figure S4B) with primers ori2_InFusion_F and ori2_InFusion_R, respectively.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
68
Materials & Methods
4.10.3 InFusion cloning to generate plasmid pACG_araBAD
Once PCR conditions were optimized for each DNA fragment, the three DNA segments were joint
together according to the InFusion cloning strategy (Figure 9) to generate plasmid pACG_araBAD
(Figure S5A).
Figure 9. InFusion cloning procedure followed to generate pACG_araBAD plasmid. Fragment 1 was amplified by
PCR from plasmid 16ABZ5NP_1934177 with primers araBAD_InFusion_F and araBAD_InFusion_R. Fragment 2
was directly amplified from plasmid pCP20 with primers cmR_InFusion_F and cmR_InFusion_R. Fragment 3 was
directly amplified from pETcoco1 with primers ori2_InFusion_F and ori2_InFusion_R, respectively. The three
DNA segments were then joint together according to the Infusion cloning strategy. Generated with Snapgene®.
The InFusion cloning product was transformed into NEB® 5-alpha Competent E. coli (High Efficiency)
cells. Some clones growing on transformation plates were selected and a miniprep was done in order
to purify plasmids. Verification of pACG_araBAD presence was performed by restriction analysis
using enzymes XhoI and MssI. Digestion products were tested by gel electrophoresis (Figure S13),
showing that all 7 tested clones contain plasmid pACG_araBAD. In order to ensure that sequence of
plasmids from positive clones was correct, some reactions were prepared for sequencing by using
primers InFusion1_seq_F1-F17 and InFusion1_seq_R1-R16 and plasmid preps as template. Sequence
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
69
of pACG_araBAD was correct for clone 1 E. coli NEB5α pACG_araBAD (Figure S53). A bacterial
glycerol stock was prepared for clone 1 E. coli NEB5α pACG_araBAD.
4.10.4 Cloning of genes into plasmid pACG_araBAD
Once pACG_araBAD plasmid was verified, the following step was to amplify by PCR from the E.coli K-
12 BW25113 genomic DNA the genes of study (thrA, ilvA, leuA, ilvBN, ilvIH and ilvC). Operon ilvGM
was amplified from E.coli K-12 BW25113 ilvG+ (generated by Sarah Charaf) genomic DNA since ilvG is
not functional in E.coli K-12 BW25113. Primers used for amplification of aforementioned genes
contained NheI and NotI restriction sites at 5’ and are summarized in Table 18. Subsequent cloning of
the amplified genes into the tunable expression plasmid pACG_araBAD was carried out by restriction
cloning using restriction enzymes NheI and NotI in order to yield plasmids pACG_araBAD_geneX
(Figure S5B-H). Plasmid pACG_araBAD_ilvBN was generated and verified by Sarah Charaf.
Table 18. Primers and templates used for amplification of target genes from E. coli genomic DNA for
subsequent NheI- and NotI-mediated cloning into plasmid pACG_araBAD.
Each cloning product was transformed into NEB® 5-alpha Competent E. coli (High Efficiency) cells.
Some clones growing on transformation plates were selected and a miniprep was done in order to
purify plasmids. Verification of plasmid presence was performed by restriction analysis as stated
inTable 19. Digestion products were tested by gel electrophoresis (Figure S14-S18).
Table 19. Restriction enzymes used for verification of pACG_araBAD plasmid variants containing target genes
(pACG_araBAD_geneX) and the corresponding expected fragment sizes.
Plasmid
Restriction enzymes
Size of resulting fragments
pACG_araBAD
XhoI, MssI
2682 bp, 4555 bp
pACG_araBAD_thrA
XhoI, NheI
3658 bp, 6022 bp
pACG_araBAD_ilvIH
XhoI, NheI
3414 bp, 6022 bp
pACG_araBAD_ilvC
XhoI, NheI
2671 bp, 6022 bp
pACG_araBAD_ilvA
XhoI, NheI
2740 bp, 6022 bp
Gene
Primers
Template
thrA
thrA_NheI_F and thrA_NotI_R
E.coli K-12 BW25113 genomic DNA
ilvA
ilvA_NheI_F and ilvA_NotI_R
E.coli K-12 BW25113 genomic DNA
leuA
leuA_NotI_F and leuA_NheI_R
E.coli K-12 BW25113 genomic DNA
ilvGM
ilvGM_NotI_F and ilvGM_NheI_R
E.coli K-12 BW25113 ilvG+ genomic DNA
ilvIH
ilvIH_NheI_F and ilvIH_NotI_R
E.coli K-12 BW25113 genomic DNA
ilvC
ilvC_NheI_F and ilvC_NotI_R
E.coli K-12 BW25113 genomic DNA
ilvBN
ilvBN_NheI_F and ilvBN_NotI_R
E.coli K-12 BW25113 genomic DNA
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
70
Materials & Methods
pACG_araBAD_leuA
NotI, MssI
3040 bp, 5749 bp
Molecular verification of plasmid pACG_araBAD_ilvGM was not performed by restriction analysis but
by colony PCR. Some clones growing on transformation plates were selected and verified by colony
PCR using primers InFusion1_seq_F4 and InFusion1_seq_R4. PCR products were tested by gel
electrophoresis (Figure S19).
Afterwards, a midiprep from clones verified by restriction analysis or PCR was carried out and some
reactions were prepared for sequencing. Forward and reverse primers were designed so that they
covered the plasmid region where the corresponding gene was cloned, hence generating overlapping
reads (Table 20).
Table 20. Sequencing primers for verification of pACG_araBAD plasmid variants containing target genes
(pACG_araBAD_geneX).
pACG_araBAD
_leuA
pACG_araBAD
_ilvC
pACG_araBAD
_ilvA
pACG_araBAD
_ilvIH
pACG_araBAD
_thrA
pACG_araBAD
_ilvGM
InFusion1_seq_
F4
InFusion1_seq_
R4
leuA_F1_seq
leuA_F2_seq
leuA_F3_seq
leuA_R1_seq
leuA_R2_seq
leuA_R3_seq
InFusion1_seq_
F4
InFusion1_seq_
R4
ilvC_F1_seq
ilvC _F2_seq
ilvC _F3_seq
ilvC _R1_seq
ilvC _R2_seq
ilvC _R3_seq
InFusion1_seq_
F4
InFusion1_seq_
R4
ilvA_F1_seq
ilvA _F2_seq
ilvA _F3_seq
ilvA _R1_seq
ilvA _R2_seq
ilvA _R3_seq
InFusion1_seq_
F4
InFusion1_seq_
R4
ilvIH_F1_seq
ilvIH _F2_seq
ilvIH _F3_seq
ilvIH _F4_seq
ilvIH _F5_seq
ilvIH _R1_seq
ilvIH _R2_seq
ilvIH _R3_seq
ilvIH _R4_seq
ilvIH _R5_seq
InFusion1_seq_
F4
InFusion1_seq_
R4
thrA_F1_seq
thrA _F2_seq
thrA _F3_seq
thrA _F4_seq
thrA _F5_seq
thrA _R1_seq
thrA _R2_seq
thrA _R3_seq
thrA _R4_seq
thrA _R5_seq
InFusion1_seq_
F4
InFusion1_seq_
R4
ilvGM_F1_seq
ilvGM_F2_seq
ilvGM_F3_seq
ilvGM_F4_seq
ilvGM_R1_seq
ilvGM_R2_seq
ilvGM_R3_seq
ilvGM_R4_seq
Plasmid sequence was correct for clone 2 E. coli NEB5α pACG_araBAD_leuA (Figure S54), clone 4 E.
coli NEB5α pACG_araBAD_ilvC (Figure S55), clone 5 E. coli NEB5α pACG_araBAD_ilvA (Figure S56),
clone 3 E. coli NEB5α pACG_araBAD_thrA (Figure S57), clone 1 E. coli NEB5α pACG_araBAD_ilvIH
(Figure S58) and clone 1 E. coli NEB5α pACG_araBAD_ilvGM (Figure S59). Bacterial glycerol stocks
were prepared for verified clones.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
71
4.11 Design of an m-toluine-based tunable expression plasmid
(pACG_XylSPm)
A tunable expression plasmid under the control of the XylS/Pm promoter system (pACG_XylSPm) was
engineered, as an alternative to plasmid pACG_araBAD, due to its numerous advantages (section
4.11.1). Genes leuA, ilvC, ilvA, ilvIH, ilvBN and thrA were afterwards cloned (section 4.11.2). Resulting
plasmids would allow expression regulation of target genes by exougenous addition of m-toluate.
4.11.1 Generation of plasmid pACG_XylSPm
Sequence of the XylS/Pm promoter region was obtained from pJB658 (Blatny et al., 1997) (Figure
S4D). This vector was proven to be successful for fine tuning expression in E.coli (Brautaset et al.,
1998; 2000; Winther-Larsen et al., 2000; Sletta et al., 2004). pJB658 was selected as a template due
to its numerous advantages compared with other vectors of its type (Blatny et al., 1997). In addition,
pJB658 contains the unmodified wild type Pm promoter variant, which demonstrated to have a less
basal expression that other mutagenized high-level expression variants (Balzer et al., 2013; Binder et
al., 2016). The region of pJB658 containing XylS and Pm was chemically synthesized and subsequently
cloned in plasmid 16ADCJKP_2028165_XylSPm (Figure S3C) by GeneArt, so that at both extremes of
the cloned XylS/Pm promoter region MssI and NheI restriction sites were introduced. The XylS/Pm
genetic region flanked by restriction sites was extracted from plasmid 16ADCJKP_2028165_XylSPm
by cutting with NheI, PstI and MssI. The digestion product was loaded into a preparative gel and the
band corresponding to the fragment containing NheI and MssI restriction sites at both ends (2174
bp) was cut and purified from the gel. In parallel, vector pACG_araBAD was also digested with NheI
and MssI. Both vector and insert were ligated (Figure 10) to generate plasmid pACG_XylSPm (Figure
S6A). The ligation product was transformed into NEB® 5-alpha Competent E. coli cells. Some clones
growing on transformation plates were selected and molecular verification of plasmid pACG_XylSPm
was performed by colony PCR using primers InFusion1_seq_F17 and InFusion1_seq_R4. PCR
products were tested by gel electrophoresis (Figure S20), showing that clones 13, 16 and 19 were
positive. Afterwards, a midiprep from PCR-verified clones was carried out and some reactions were
prepared for sequencing by using primers InFusion1_seq_F17 and InFusion1_seq_R4, which covered
the plasmid region where the XylS/Pm promoter was cloned. Sequence of plasmid pACG_XylSPm
from clone 16 E. coli NEB5α pACG_XylSPm was successfully verified (Figure S60) and a bacterial
glycerol stock was prepared.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
72
Materials & Methods
Figure 10. Cloning procedure followed to generate pACG_XylSPm plasmid. Generated with Snapgene®.
4.11.2 Cloning of genes into plasmid pACG_XylSPm
Once pACG_XylSPm plasmid was verified, the following step was to amplify by PCR from the E.coli K-
12 BW25113 genomic DNA the genes of study (thrA, ilvA, leuA, ilvIH, ilvBN and ilvC). Primers used for
amplification of aforementioned genes contained NheI and NotI restriction sites at 5’ and are
summarized in Table 18. Subsequent cloning of the amplified genes into the tunable expression
plasmid pACG_XylSPm was carried out by restriction cloning using restriction enzymes NheI and NotI
in order to yield plasmids pACG_XylSPm_geneX (Figure S6B-G). The cloning product was transformed
into NEB® 5-alpha Competent E. coli cells. Molecular verification of the corresponding
pACG_XylSPm_geneX plasmid was carried out by colony PCR. Some clones growing on
transformation plates were selected and verified by colony PCR using primers InFusion1_seq_F17
and InFusion1_seq_R4. PCR products were tested by gel electrophoresis (Figure S21-S26). In order to
ensure that sequence of plasmids from positive clones was correct, some reactions were prepared
for sequencing by using colony PCR products as template and primers listed in Table 21.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
73
Table 21. Sequencing primers for verification of pACG_XylSPm plasmid variants containing target genes
(pACG_XylSPm_geneX).
pACG_XylSPm
_leuA
pACG_ XylSPm
_ilvC
pACG_ XylSPm
_ilvA
pACG_ XylSPm
_ilvIH
pACG_ XylSPm
_thrA
pACG_ XylSPm
_ilvBN
InFusion1
_seq_R4
leuA_R1_seq
InFusion1
_seq_R4
ilvC _R3_seq
InFusion1
_seq_R4
ilvA _R3_seq
InFusion1
_seq_R4
ilvIH_R5_seq
InFusion1
_seq_R4
thrA _R5_seq
InFusion1
_seq_R4
ilvBN_R2_seq
ilvBN_F1_seq
Plasmid sequence was correct for clone 4 E. coli NEB5α pACG_XylSPm_leuA (Figure S61), clones 1, 3,
6 and 9 E. coli NEB5α pACG_XylSPm_ilvC (Figure S62), clone 21 E. coli NEB5α pACG_XylSPm_ilvIH
(Figure S63), clones 2 and 8 E. coli NEB5α pACG_XylSPm_thrA (Figure S64), clone 3 E. coli NEB5α
pACG_XylSPm_ilvBN (Figure S65) and clones 4 and 6 E. coli NEB5α pACG_XylSPm_ilvA (Figure S66)
Bacterial glycerol stocks were prepared for verified clones.
4.12 Development of strains to regulate expression of the BCAA biosynthesic
genes (geneX-tunable E. coli)
First of all, single knock-outs of the target endogenous genes involved in the BCAA synthesis were
performed (sections 4.12.1 and 4.12.2). The respective electro-competent E. coli K-12 BW25113
ΔgeneX mutants (geneX: leuA, thrA, ilvA, ilvC, ilvIH, ilvBN or ilvGM) were transformed with
pSW3_lacI+ (section 4.12.3), a high copy plasmid encoding mini-proinsulin, necessary for evaluation
of ncBCAA mis-incorporation; and a tunable expression plasmid variant pACG_araBAD_geneX
(section 4.12.4) or pACG_XylSPm_geneX (section 4.12.5), a 1-copy plasmid allowing expression
regulation of target genes, previously knocked-out (Figure 11).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
74
Materials & Methods
Figure 11. Overview of the genetic modifications carried out in this study in order to generate geneX-tunable E.
coli strains. Each engineered geneX-tunable E. coli lacked a certain endogenous target gene, generally indicated
as geneX in this work (E. coli K-12 BW25113 ΔgeneX), and contained one plasmid allowing expression
regulation of such target gene by exogenous L-arabinose (pACG_araBAD_geneX) and one plasmid enabling
IPTG-mediated expression of recombinant mini-proinsulin for subsequent analysis of the impurity profile
(pSW3_lacI+). A total of seven tunable E. coli strains were engineered, one for each target gene: leuA, thrA, ilvA,
ilvC, ilvIH, ilvBN or ilvGM.
4.12.1 Generation of strains E. coli K-12 BW25113 ΔgeneX (geneX: leuA, thrA, ilvA and
ilvC)
Strain E. coli K-12 BW25113 as well as single knock-out mutants E. coli K-12 BW25113 leuA:kanR, E.
coli K-12 BW25113 thrA:kanR, E. coli K-12 BW25113 ilvA:kanR and E. coli K-12 BW25113 ilvC:kanR
containing pKD46 were acquired from the E. coli Genetic Stock Center (CGSC) from Yale University.
Those mutant strains belong to the so-called KEIO collection (Baba et al., 2006). The genome of those
mutant strains contains a kanamycin resistance marker substituting the target gene. CGSC
identification for each acquired strain is indicated in Table S5.
Plasmid pKD46 was curated from the acquired E. coli K-12 BW25113 single knock-out mutants
(section 4.12.1.1) and the respective electro-competent cells were transformed with pCP20, a
temperature-sensitive plasmid encoding a flipase, in order to trigger removal of the kanamycin
cassette from the genome by FRT-specific recombination (sections 4.12.1.2 and 4.12.1.3). Plasmid
pCP20 was then curated and removal of the antibiotic resistance marker from E. coli K-12 BW25113
single mutants was tested by colony PCR (section 4.12.1.4).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
75
4.12.1.1 Curation of pKD46
Curation of plasmid pKD46 was performed according to section 4.7.6 (plates not shown). In order to
molecularly confirm removal of plasmid pKD46, one of the clones growing in LB and LB + kanamycin
plates was selected for each E. coli K-12 BW25113 geneX:kanR tested strain and a miniprep was
done in order to purify plasmids. Verification of pKD46 presence was performed by restriction
analysis using EcoRI. Digestion products were tested by gel electrophoresis (Figure S27), showing that
all tested clones lacked pKD46.
4.12.1.2 Transformation of E. coli K-12 BW25113 geneX:kanR with pCP20 and removal of the
kanamycin resistant cassette
A miniprep of an E. coli BT340 (CGSC#7629) culture was done in order to isolate plasmid pCP20.
Electro-competent E. coli K-12 BW25113 geneX:kanR cells were prepared and transformed with
plasmid pCP20 by electroporation. In order to molecularly confirm transformation of plasmid pCP20,
2 clones growing in transformation plates were selected for each E. coli K-12 BW25113 geneX:kanR
tested strain (for E. coli K-12 BW25113 ilvI:kanR 4 clones were selected) and a miniprep was done in
order to purify plasmids. Verification of pCP20 presence was performed by restriction analysis using
EcoRI. Digestion products were tested by gel electrophoresis (Figure S28), showing that all tested
clones contained pCP20.
4.12.1.3 Verification of kanamycin-resistance cassette removal from genomic DNA of E. coli K-
12 BW25113 ΔgeneX mutants
In order to molecularly confirm the flippase-mediated removal of kanamycin-resistance cassette
from genomic DNA in E. coli K-12 BW25113 ΔgeneX mutants containing pCP20, 2 clones were
selected for each E. coli K-12 BW25113 ΔgeneX strain (for E. coli K-12 BW25113 ΔilvI 3 clones were
selected) and were subjected to colony PCR using primers listed in Table 22. PCR products were
tested by gel electrophoresis (Figure S29), showing that for all tested clones kanamycin resistance
marker was successfully removed from the genome.
Table 22. Primers used for PCR-verification of flippase-mediated removal of the kanamycin resistance cassette
from the genome of strains E. coli K-12 BW25113 ΔgeneX pCP20 and the corresponding expected PCR product
sizes for both kanS and kanR variants.
Strain to verify
Primers
Expected size PCR product (bp)
E. coli K-12 BW25113 ΔthrB pCP20
thrB_F, thrB_R
kanS1: 1081, kanR2: 2281
E. coli K-12 BW25113 ΔthrC pCP20
thrC_F, thrC_R
kanS: 691, kanR: 1891
E. coli K-12 BW25113 ΔilvA pCP20
ilvA_F, ilvA_R
kanS: 727, kanR: 1927
E. coli K-12 BW25113 ΔilvB pCP20
ilvB_F, ilvB_R
kanS: 1307, kanR: 2507
E. coli K-12 BW25113 ΔthrA pCP20
thrA_F, thrA_R
kanS: 1164, kanR: 2364
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
76
Materials & Methods
E. coli K-12 BW25113 ΔleuA pCP20
leuA_F, leuA_R
kanS: 746, kanR: 1946
E. coli K-12 BW25113 ΔilvC pCP20
ilvC_F, ilvC_R
kanS: 706, kanR: 1906
E. coli K-12 BW25113 ΔilvI pCP20
ilvIH_F, ilvIH_R
kanS: 1246, kanR: 2446, wild type:
2844
1: kan-sensitive clone (kanS), i.e. clone were kan resistance marker was removed from the genome.
2: kan-resistant clone (kanR), i.e. clone were kan resistance marker still remains integrated into the genome.
4.12.1.4 Curation of pCP20
Curation of plasmid pCP20 was performed according to section 4.7.6. For strains E. coli K-12
BW25113 ΔleuA, ΔilvC and ΔilvI a second temperature treatment of 42 °C was necessary to
completely remove plasmid pCP20. Plate results are shown in Figure S67. A bacterial glycerol stock
was prepared for clone 2.3.1 E. coli K-12 BW25113 ΔleuA, clone 2.2.1 E. coli K-12 BW25113 ΔilvC,
clone 1.2.1 E. coli K-12 BW25113 ΔilvI, clone 1.1 E. coli K-12 BW25113 ΔilvA, clone 1.1 E. coli K-12
BW25113 ΔilvB, clone 1.1 E. coli K-12 BW25113 ΔthrA, clone 1.1 E. coli K-12 BW25113 ΔthrB and
clone 1.1 E. coli K-12 BW25113 ΔthrC.
4.12.2 Generation of strains E. coli K-12 BW25113 ΔilvIH and E. coli K-12 BW25113 ΔilvBN
The knock-out strains E. coli K-12 BW25113 ΔilvIH and ΔilvBN were not acquired from CGSC but
generated in the laboratory. Strain E. coli K-12 BW25113 ΔilvIH was generated in the context of this
thesis while strain E. coli K-12 BW25113 ΔilvBN was generated by Sarah Charaf. A summary of the
procedure carried out to generate strain E. coli K-12 BW25113 ΔilvIH is described below. Electro-
competent E. coli K-12 BW25113 cells were transformed with pKD46, a temperature-sensitive helper
plasmid, expressing the elements enabling homologous recombination (section 4.12.2.1). Knock-out
mutants for the operons ilvIH and ilvBN were then generated by transformation of electro-
competent E. coli K-12 BW25113 cells containing pKD46 with the respective deletion cassette,
previously obtained by PCR from pKD3 or pKD4. PCR-based verification was carried out to test proper
integration of the deletion cassette into the genome (section 4.12.2.2). Plasmid pKD46 was curated
(section 4.12.2.3) and the respective electro-competent E. coli K-12 BW25113 mutants were
transformed with pCP20, a temperature-sensitive plasmid encoding a flipase, in order to trigger
removal of the antibiotic cassette from the genome by FRT-specific recombination (section 4.12.2.4).
Plasmid pCP20 was curated and removal of the antibiotic resistance marker from E. coli K-12
BW25113 ΔilvIH and ΔilvBN single mutants was tested by sequencing (section 4.12.2.5).
4.12.2.1 Transformation of E. coli K-12 BW25113 with pKD46
Electro-competent E. coli K-12 BW25113 cells were prepared and transformed with plasmid pKD46
by electroporation. Six clones growing on transformation plates were selected and a miniprep was
done in order to purify plasmids. Verification of pKD46 presence was performed by restriction
analysis using EcoRI. Digestion products were tested by gel electrophoresis (Figure S30), showing that
all 6 tested clones contained pKD46.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
77
4.12.2.2 Transformation of E. coli K-12 BW25113 pKD46 with the ilvIH-deletion cassette
The ilvIH knock-out cassette was generated by PCR amplification of the chloramphenicol or
kanamycin resistance cassette with primers KO3_F and KO3_R, by using as template plasmid pKD3 or
pKD4, respectively. PCR primers were designed so that they contained 36 bp homologous to the
genomic region flanking operon ilvIH at 5’ and 20 bp homologous to priming sites 1 and 2 of the
template plasmids at 3’ (Table 23). PCR was performed with Phusion DNA Polymerase by using
different buffer combinations (buffer HF or GC plus DMSO) and PCR products were tested by gel
electrophoresis (Figure S31).
Table 23. Primers used to generate the ilvIH knock-out cassette. Highlighted in yellow at 5’ are 36 bp
homologous to the genomic region flanking operon ilvIH. At 3’, 20 bp homologous to priming sites 1 and 2
present in plasmids pKD3 and pKD4 are highlighted in blue and green respectively.
Primer name
Primer sequence (5’ – 3’)
KO3_F
TTTACACATTTTTTCCGTCAAACAGTGAGGCAGGCCGTGTAGGCTGGAGCTGCTTC
KO3_R
ACATGTTGGGCTGTAAATTGCGCATTGAGATCATTCATGGGAATTAGCCATGGTCC
A preparative gel was then prepared and 200 μL of the PCR product generated by using buffer GC
plus DMSO and pKD3 as template were loaded. After electrophoresis, the corresponding band of
1105 bp was cut from the gel, DNA was purified and concentration was determined by Nanodrop.
Purified PCR product was additionally treated with DpnI in order to remove any rest of plasmid DNA
template. Electro-competent E. coli K-12 BW25113 pKD46 cells were prepared and transformed with
the purified ilvIH knock-out cassette by electroporation. In order to molecularly confirm ilvIH knock
out, genomic DNA of 10 colonies grown on the transformation plates was extracted and PCR was
carried out with primers ilvIH_F and ilvIH_R. PCR products were tested by gel electrophoresis (Figure
S32), being clone 1 the only positive clone.
4.12.2.3 Curation of pKD46
Curation of plasmid pKD46 from clone 1 E. coli K-12 BW25113 ilvIH:cmR pKD46 was performed
according to section 4.7.6. In this case, a second temperature treatment of 42 °C was necessary to
completely remove plasmid pKD46. Plate results are shown in Figure S68 and indicated that all
potential subclones originating from clone 1 E. coli K-12 BW25113 ilvIH:cmR pKD46 lost plasmid
pKD46 after temperature treatment.
4.12.2.4 Transformation of E. coli K-12 BW25113 ilvIH:cmR with pCP20 and removal of the
chloramphenicol resistant cassette
Electro-competent E. coli K-12 BW25113 ilvIH:cmR cells were prepared and transformed with
plasmid pCP20 by electroporation. In order to molecularly confirm if flippase-mediated site specific
recombination took place and chloramphenicol resistant marker was removed from the genome,
genomic DNA of 4 colonies grown on the transformation plates was extracted and PCR was carried
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
78
Materials & Methods
out with primers ilvIH_F2 and ilvIH_R. PCR products were tested by gel electrophoresis (Figure S33),
showing that all evaluated clones were positive.
4.12.2.5 Curation of pCP20
Curation of plasmid pCP20 was performed according to section 4.7.6. In this case, a second
temperature treatment of 42 °C was necessary to completely remove plasmid pCP20. Plate results
are shown in Figure S69, concluding that all potential clones lost plasmid pCP20.
In order to ensure that E. coli K-12 BW25113 ΔilvIH mutant was correct, sequencing of the mutated
region was performed. Hence, a PCR of the genomic region flanking the ilvIH deletion was carried out
with primers ilvIH_F2 and ilvIH_R and the resulting PCR product was sent for sequencing. Sequencing
reactions were done with the same primers as for PCR. According to sequencing results, the
sequence on the ilvIH-deleted region in E. coli K-12 BW25113 ΔilvIH was:
tttcttttcacctttcctcctgtttattcttattaccccgtgtttatgtctctggctgccaattgcttaagcaagatcggacggttaatgtgttttacacat
tttttccgtcaaacagtgaggcaggccgtgtaggctggagctgcttcgaagttcctatactttctagagaataggaacttcggaataggaactaa
ggaggatattcatatggaccatggctaattcccatgaatgatctcaatgcgcaatttacagcccaacatgtcacgttgggctttttttgcgaaatca
gtgggaacctggaataaaagcagttgccgcagttaattttctgcgcttagatgttaatgaatt. Highlighted in green appears the
sequence located upstream from the ilvIH deletion while in red appears the sequence located
downstream. The scar segment is displayed in yellow, with the FRT site in bold. After sequence
verification, a bacterial glycerol stock of strain E. coli K-12 BW25113 ΔilvIH was prepared.
4.12.3 Transformation of pSW3_lacI+ into E. coli K-12 BW25113 ΔgeneX
Electro-competent E. coli K-12 BW25113 ΔgeneX cells were prepared and transformed with the
previously verified plasmid pSW3_lacI+ by electroporation. In order to molecularly confirm presence
of plasmid pSW3_lacI+, some clones growing on transformation plates were selected for each E. coli
K-12 BW25113 ΔgeneX tested strain and were subjected to colony PCR using primers pSW3_R3_seq
and pSW3_R6_seq. PCR products were tested by gel electrophoresis (Figure S34-S37). According to
the gels all clones were positive, being confirmed that they contained plasmid pSW3_lacI+. A
bacterial glycerol stock was prepared with a verified clone from each tested strain.
4.12.4 Transformation of pACG_araBAD_geneX into E. coli K-12 BW25113 ΔgeneX
pSW3_lacI+ (geneX-tunable E. coli)
Electro-competent E. coli K-12 BW25113 ΔgeneX pSW3_lacI+ cells were prepared and transformed
with the previously verified plasmid pACG_araBAD_geneX by electroporation. In order to molecularly
confirm presence of plasmid pACG_araBAD_geneX, some clones growing on transformation plates
were selected for each E. coli K-12 BW25113 ΔgeneX pSW3_lacI+ tested strain and were subjected to
colony PCR using primers InFusion1_seq_F4 and InFusion1_seq_R4. PCR products were tested by gel
electrophoresis (Figure S38-S44). According to the gels all clones were positive, being confirmed that
they contained plasmid pACG_araBAD_geneX. A bacterial glycerol stock was prepared with a verified
clone from each tested strain. The generated strains would allow transcription regulation of single
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Materials & Methods
79
BCAA biosynthesic genes by exogenous L-arabinose addition. Through this dissertation the term
geneX-tunable E. coli is often used instead of the complete strain name (E. coli K-12 BW25113
ΔgeneX pSW3_lacI+ pACG_araBAD_geneX) for simplification reasons.
4.12.5 Transformation of pACG_XylSPm_geneX into E. coli K-12 BW25113 ΔgeneX
pSW3_lacI+
Electro-competent E. coli K-12 BW25113 ΔgeneX pSW3_lacI+ cells were prepared and transformed
with the previously verified plasmid pACG_XylSPm_geneX by electroporation. In order to molecularly
confirm presence of plasmid pACG_XylSPm_geneX, some clones growing on transformation plates
were selected for each E. coli K-12 BW25113 ΔgeneX tested strain and were subjected to colony PCR
using primers InFusion1_seq_F17 and InFusion1_seq_R4. PCR products were tested by gel
electrophoresis (Figure S45-S49). According to the gels all clones were positive, being confirmed that
they contained plasmid pACG_XylSPm_geneX. A bacterial glycerol stock was prepared with a verified
clone from each tested strain. The generated strains would allow transcription regulation of single
BCAA biosynthesic genes by exogenous m-toluate addition.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
80
Results
5. Results
5.1 Analysis of mini-proinsulin expression in E. coli K-12 W3110M and E. coli
K-12 BW25113 containing plasmid pSW3
Together with the novel Pm/XylS promoter, the classic araBAD promoter was also utilized in this
thesis in order to regulate expression of target genes involved in the BCAA biosynthetic pathway. In
order to ensure functionality of the araBAD promoter, the host strain must be deficient in arabinose
catabolizing enzymes. For this reason, strain E. coli K-12 BW25113 was selected as the model
organism for this study, since it contains a deletion of the endogenous araBAD operon. However, the
plasmid expressing recombinant mini-proinsulin under the control of a Ptac promoter selected for this
study (pSW3) was just previously tested on strain E. coli K-12 W3110M, showing here good levels of
expression and low promoter leakiness. Hence, it was necessary to evaluate first performance of
plasmid pSW3 in the background of E. coli K-12 BW25113.
The aim of this experiment was to evaluate expression of recombinant mini-proinsulin in both strains
E. coli K-12 W3110M and E. coli K-12 BW25113 containing plasmid pSW3 by SDS-PAGE. It was aimed
to check how leaky was the expression of the Ptac promoter by using two different E. coli K-12 genetic
backgrounds (W3110M and BW25113) and to confirm if, unlike E. coli K-12 W3110M (lacIq), E. coli K-
12 BW25113 is a lacI+ strain.
5.1.1 Cultivation conditions
Pre-culture was prepared as follows: 10 mL of M9 medium were introduced in sterile 100 mL
Erlenmeyer shake flasks. One flask was prepared for each tested strain (E. coli K-12 W3110M, E. coli
K-12 W3110M pSW3, E. coli K-12 BW25113 and E. coli K-12 BW25113 pSW3). Ampicillin was
supplemented to a final concentration of 100 µg/mL for strains containing pSW3. Fresh cells were
inoculated into the medium and cultures were incubated at 37 °C und 220 rpm, overnight. The main
culture was then prepared and cultivated as follows: 8 sterile 300 mL Erlenmeyer shake flasks were
filled with 30 mL M9 medium. 2 shake flasks were prepared for each tested strain. OD600nm of pre-
cultures was measured and the volume from the pre-culture needed to generate a main culture at
0.05 OD600nm in a final volume of 30 mL was calculated. The corresponding calculated volume of pre-
culture needed was introduced in the 300 mL sterile Erlenmeyer shake flasks. Ampicillin was
supplemented to a final concentration of 100 µg/mL for strains containing pSW3. Resulting cultures
were incubated at 37°C and 220 rpm until OD600nm reached 0.5-0.8. Then, 300 μL of 40% glucose were
added to all cultures while IPTG was added to a concentration of 0.5 mM only to one of the shake
flasks available for each tested strain (induced sample). The other shake flask was used as negative
control for induction (non-induced sample). Cultures were then left incubating at 37°C and 220 rpm.
Around 20h after induction, OD600nm was measured and the corresponding volume needed to
generate a 1 mL solution at OD600nm =1 was taken out, centrifuged (6000g, 5min) and the resulting
pellet was stored at -20°C for SDS-PAGE analysis.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
81
5.1.2 Mini-proinsulin analysis with SDS-PAGE
Mini-proinsulin expression was evaluated by SDS-PAGE with samples taken 20 h after IPTG induction.
For each tested strain, 2 cultivations were evaluated: one with IPTG induction and the other, without
induction (Figure 12).
Figure 12. SDS-PAGE showing intracellular protein profile for different E. coli strains (E. coli K-12 W3110M, E.
coli K-12 W3110M pSW3, E. coli K-12 BW25113 and E. coli K-12 BW25113 pSW3), grown on M9 medium, about
20h after induction of protein expression with IPTG (20hAI). Intracellular protein profiles of all tested strains
and conditions correspond to an OD600nm of 0.5, so that gel lanes are comparable. Plasmid pSW3 expresses an
11 kDa recombinant mini-proinsulin, indicated by the red arrow. For each tested strain, an IPTG-induced
sample (+I) as well as a non-induced sample (-I) is included. Std refers to the SeeBlue® Plus2 Pre-stained Protein
Standard. Molecular weights of the protein standard are shown in kDa.
As expected, for the control empty strains E. coli K-12 W3110M and E. coli K-12 BW25113 (not
containing plasmid pSW3) no mini-proinsulin expression was reported for both induced and non-
induced samples (Figure 12: lanes 2-3 and 6-7, respectively). Strain E. coli K-12 W3110M pSW3
showed good levels of mini-proinsulin expression 20 h after IPTG induction as well as really low
promoter leakiness, since no mini-proinsulin expression was reported for the non-induced sample
(Figure 12: lanes 4 and 5). However, as opposed to E. coli K-12 W3110M, Ptac promoter controlling
expression of mini-proinsulin in plasmid pSW3 was leaky in the genetic background of E. coli K-12
BW25113, since there was no significant difference in reported mini-proinsulin expression levels
between the induced and non-induced samples (Figure 12: lanes 8 and 9). In addition, expression
intensity achieved by strain E. coli K-12 BW25113 pSW3 was clearly weaker than the one obtained for
strain E. coli K-12 W3110M pSW3 (Figure 12: lanes 5 and 9). Hence, alternative pSW3 variants
expressing different levels of LacI repressor were constructed and tested in the genetic background
of E. coli K-12 BW25113 in order to avoid promoter leakiness and improve recombinant protein
production.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
82
Results
5.2 Analysis of mini-proinsulin expression in E. coli K-12 BW25113 containing
different variants of plasmid pSW3
The aim of this experiment was to compare the mini-proinsulin expression level as well as the
leakiness of Ptac promoter by SDS-PAGE in cultures of E. coli K-12 BW25113 containing 3 different
variants of plasmid pSW3 (pSW3, pSW3_lacI+ and pSW3_lacIq), each of them expressing different
amounts of LacI repressor. Strain E. coli K-12 W3110M expressing pSW3 was used as a reference for
comparison since it previously showed optimal expression levels for recombinant mini-proinsulin
(see section 5.1.2).
5.2.1 Cultivation conditions
Pre-culture was prepared as follows: 20 mL of modified Davis and Mingioli medium were introduced
in sterile 100 mL Erlenmeyer shake flasks. One flask was prepared for each tested strain (E. coli K-12
W3110M pSW3, E. coli K-12 BW25113 pSW3, E. coli K-12 BW25113 pSW3_lacI+ and E. coli K-12
BW25113 pSW3_lacIq). Ampicillin was supplemented to a final concentration of 100 µg/mL. Fresh
cells were inoculated into the medium and cultures were incubated at 37 °C und 220 rpm, overnight.
The main culture was prepared and cultivated as described as follows: 8 sterile 300 mL Erlenmeyer
shake flasks were filled with 30 mL of modified Davis and Mingioli medium. 2 shake flasks were
prepared for each tested strain. OD600nm of pre-cultures was measured and the volume from the pre-
culture needed to generate a main culture at 0.1 OD600nm in a final volume of 30 mL was calculated.
The corresponding calculated volume of pre-culture needed was introduced in the 300 mL sterile
Erlenmeyer shake flasks. Ampicillin was supplemented to a final concentration of 100 µg/mL.
Resulting cultures were incubated at 37°C and 220 rpm until OD600nm reached 0.3-0.4. Then, 300 μL of
40% glucose were added to all cultures while IPTG was added to a concentration of 0.5 mM only to
one of the shake flasks available for each tested strain (induced sample). The other shake flask was
used as negative control for induction (non-induced sample). Cultures were then left incubating at
37°C and 220 rpm. Short before induction and 1h, 3h, 6h and overnight after induction, OD600nm was
measured and and the corresponding volume needed to generate a 1 mL solution at OD600nm =1 was
taken out, centrifuged (10 000 g, 5 min) and the resulting pellet was stored at -20°C for SDS-PAGE
analysis.
5.2.2 Mini-proinsulin analysis with SDS-PAGE
For each tested strain, mini-proinsulin expression was evaluated by SDS-PAGE with samples taken
over cultivation time (1h, 3h, 6h and overnight after induction). In addition, for each strain, 2
cultivations were evaluated: one with IPTG induction and the other, without induction (Figure 13).
Calculation of the relative quantity (Rel. q.) of mini-proinsulin was performed with the software
Image Lab (Biorad) as follows: intensity of the band containing mini-proinsulin (band around 11 kDa
present in induced samples in the SDS-PAGE gels) was determined for each lane. Intensity of the first
band containing mini-proinsulin (band around 11 kDa present in lane 3 of the SDS-PAGE gels) was
used as a reference for the calculation of the relative quantity of mini-proinsulin in the other bands.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
83
For instance, the relative quantity of mini-proinsulin in a certain lane would be equal to the intensity
of the mini-proinsulin band in such lane divided by the intensity of that band in lane 3.
Figure 13. SDS-PAGE showing intracellular protein profile for different E. coli strains (E. coli K-12 W3110M
pSW3, E. coli K-12 BW25113 pSW3, E. coli K-12 BW25113 pSW3_lacI+ and E. coli K-12 BW25113 pSW3_lacIq),
grown on modified Davis & Mingioli medium, 1h (A), 3h (B), 6h (C) and overnight (D) after induction of protein
expression with IPTG (onAI). Intracellular protein profiles of all tested strains and conditions correspond to an
OD600nm of 0.5, so that gel lanes are comparable. Plasmid pSW3 and its variants express an 11 kDa recombinant
mini-proinsulin, indicated by the red arrow. For each tested strain, an IPTG-induced sample (+I) as well as a
non-induced sample (-I) is included. Std refers to the SeeBlue® Plus2 Pre-stained Protein Standard. Molecular
weights of the protein standard are shown in kDa. Relative quantity (Rel. q.) of mini-proinsulin is also indicated.
According to the results, E. coli K-12 BW25113 (lacI+) expressing pSW3_lacI+ presents a similar
induction behavior than E. coli K-12 W3110M (lacIq) expressing pSW3 (Figure 13: lanes 2-3 and 6-7).
E. coli K-12 BW25113 expressing pSW3 did not show any expression after IPTG induction (Figure 13:
A B
Rel.q. 1 1.03 0.61
Rel.q. 1 0.94 0.68
Rel.q. 1 1 0.64
Rel.q. 1 0.9 0.61
C D
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
84
Results
lane 5). However, in section 5.1.2, protein production was reported for that strain 20 h after
induction. This incongruity is explained due to an experimental error, since strain employed in this
study was not actually E. coli K-12 BW25113 expressing pSW3 but the empty E. coli K-12 BW25113
strain. Strain E. coli K-12 BW25113 (lacI+) expressing pSW3_lacIq also achieved protein expression
(Figure 13: lane 9). However, expression reported around a 40 % decrease than in E. coli K-12
W3110M pSW3 and E. coli K-12 BW25113 pSW3_lacI+ (Figure 13: lanes 3, 7 and 9). In conclusion, it
seems that for the genetic background of E. coli K-12 BW25113 (lacI+) the pSW3 plasmid variant
containing the lacI+ genetic region is the optimal one since it shown a similar induction behavior than
E. coli K-12 W3110M (lacIq) pSW3, thus achieving a good production of mini-proinsulin. In addition,
Ptac promoter seems to be quite good regulated since, in the absence of IPTG, no production was
achieved (Figure 13: lane 6). Hence, strain E. coli K-12 BW25113 (lacI+) pSW3_lacI+ was selected as
the model organism for further investigation in the current thesis.
5.3 Evaluation of L-arabinose induction in E. coli BW25113 ΔgeneX
expressing pSW3_lacI+ and pACG_araBAD_geneX (geneX-tunable E. coli)
The aim of this experiment was to evaluate the effect of using different concentrations of L-
arabinose in the expression level of target genes involved in the BCAA biosynthetic pathway (gene X),
which are under the control of the araBAD promoter in strains E. coli K-12 BW25113 ΔgeneX
expressing pSW3_lacI+ and pACG_araBAD_geneX (geneX-tunable E. coli).
5.3.1 Cultivation conditions
Pre-cultures were prepared as follows: 5 mL of 1:3 TUB medium were introduced in sterile 15 mL
tubes. Various pre-cultures were prepared for each tested strain. Ampicillin and chloramphenicol
were supplemented to a final concentration of 100 µg/mL and 25 μg/mL, respectively.
Chloramphenicol was not added in cultures containing strain E. coli K-12 BW25113 pSW3_lacI+.
Different concentrations of L-arabinose were added to the different pre-cultures. 50 μL of the
corresponding criostock were then inoculated into the medium and cultures were incubated at 37 °C
and 250 rpm, overnight. OD600nm was measured at 16h cultivation time.
5.3.2 Gene expression analysis
Most of the target genes investigated in this study are crucial for E. coli metabolism so that when no
L-arabinose is added, target gene is not expressed and this might have a negative effect on E. coli
growth behavior. Hence, in order to indirectly proof effectivity of L-arabinose as inducer of gene
expression in plasmids pACG_araBAD_geneX, OD600nm from the different pre-cultures was measured
after 16h cultivation. Growth behavior (OD600nm) was then plotted for every tested tunable E. coli
strain under different L-arabinose concentrations and it was compared to the non-modified strain E.
coli K-12 BW25113 pSW3_lacI+ (
Figure 14).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
85
Figure 14. Chart showing measured OD600nm after 16h cultivation of the different E. coli K-12 BW25113 ΔgeneX
pSW3_lacI+ pACG_araBAD_geneX strains under different L-arabinose concentrations. The non-modified strain
E. coli K-12 BW25113 pSW3_lacI+ is also included as a control for comparison.
For strain E. coli K-12 BW25113 ΔilvA pSW3_lacI+ pACG_araBAD_ilvA (simplified, ilvA-tunable E. coli),
L-arabinose induction showed a positive growth effect when using at least 0.025 % L-arabinose.
Growth behavior recovered levels of the control E. coli strain (E. coli K-12 BW25113 pSW3_lacI+)
when at least 0.05 % L-arabinose was employed. For ilvIH-, ilvBN- and ilvGM-tunable E. coli strains
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
86
Results
there were no significant differences in growth behavior reported between non-induced and L-
arabinose-induced samples, independently on the L-arabinose concentration employed. In those
cases, induction efficiency of L-arabinose could not be properly evaluated by monitoring cell growth.
L-arabinose induction showed a positive growth effect for leuA-tunable E. coli when using at least
0.05 % L-arabinose. Growth behavior recovered levels of the control E. coli strain when at least 0.1 %
L-arabinose was employed. L-arabinose was also proven to function for ilvC-tunable E. coli. A positive
growth effect was reported when using at least 0.05 % L-arabinose. Growth behavior approached to
levels of the control E. coli strain when at least 0.4 % L-arabinose was employed. Finally, L-arabinose
induction was also demonstrated to success for thrA-tunable E. coli. When employing at least 0.4 %
L-arabinose a positive growth effect was reported, approaching to levels of the control E. coli strain.
It could be concluded that the effective induction range of L-arabinose for plasmid variants
pACG_araBAD_geneX was between 0.05 and 1.6 % L-arabinose. Furthermore, the maximum tested
L-arabinose concentration (1.6 %) did not show a cellular toxic effect, since growth behavior
remained unaltered. It is also noteworthy that, for each target gene, different induction strength, i.e.
L-arabinose concentration, was necessary in order to trigger genetic expression levels enough to
recover cell growth levels of the control E. coli strain.
5.4 Evaluation of m-toluate induction in E. coli BW25113 ΔgeneX expressing
pSW3_lacI+ and pACG_XylSPm_geneX (geneX-tunable E. coli)
Analog to section 5.3, the aim of this experiment was to evaluate the effect of employing different
concentrations of m-toluate in the expression of target genes involved in the BCAA biosynthetic
pathway (gene X), which are under the control of the XylS/Pm promoter in strains E. coli K-12
BW25113 ΔgeneX expressing pSW3_lacI+ and pACG_XylSPm_geneX (geneX-tunable E. coli).
5.4.1 Cultivation conditions
The cultivation procedure was as described in section 5.3.1, with the following alterations: instead of
1:3 TUB medium, modified Davis & Mingioli medium was employed; instead of L-arabinose, different
concentrations of m-toluate were added to the pre-cultures.
5.4.2 Gene expression analysis
Similarly to section 5.3.2, OD600nm was plotted for every tested tunable E. coli strain under different
m-toluate concentrations and it was compared to the non-modified strain E. coli K-12 BW25113
pSW3_lacI+ (Figure 15).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
87
Figure 15. Chart showing measured OD600nm after 16h cultivation of the different E. coli K-12 BW25113 ΔgeneX
pSW3_lacI+ pACG_XylSPm_geneX strains under different L-arabinose concentrations. The non-modified strain
E. coli K-12 BW25113 pSW3_lacI+ is also included as a control for comparison.
Induction of plasmid variants pACG_XylSPm_geneX with m-toluate was not successfull since addition
of m-toluate did not show a positive effect on growth behavior if compared with the non-induced
case for any tested mutant strain. In addition, it can be concluded that 10 mM m-toluic acid triggers
cellular toxicity since under this concentration OD600nm reached almost 0 for all tested strains.
In accordance to results obtained in sections 5.3.2 and 5.4.2, the plasmid variant
pACG_araBAD_geneX was selected in this thesis for further expression regulation of target genes
involved in the BCAA biosynthetic pathway. The plasmid variant pACG_XylSPm_geneX was discarded.
5.5 Establishment of a GC-FID method allowing analysis of canonical and
non-canonical amino acids
The aim of this experiment was to establish and validate the use of the EZ:faastTM free (physiological)
amino acid analysis kit for canonical and non-canonical amino acid analysis by gas chromatography-
flame ionization detection (GC-FID). To achieve that, following activities were carried out:
determination of retention times and elaboration of calibration curves for each amino acid (sections
5.5.1 and 5.5.2), evaluation of the hydrolysis effect on amino acid analysis (section 5.5.3) and
validation of the method by analyzing a pure protein (section 5.5.4).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
88
Results
5.5.1 Elaboration of calibration curves and determination of retention times for ncBCAA
norvaline, norleucine and β-methylnorleucine
The following amino acid mixtures were prepared for calibration: calibration level 1 (25 µM): 12.5 μL
200 μM ncBCAA plus 100 μL 200 μM ABA; calibration level 2 (50 µM): 25 μL 200 μM ncBCAA plus 100
μL 200 μM ABA; calibration level 3 (100 µM): 50 μL 200 μM ncBCAA plus 100 μL 200 μM ABA;
calibration level 4 (200 µM): 100 μL 200 μM ncBCAA plus 100 μL 200 μM ABA. Each mixture was used
for amino acid analysis according to sections 4.8.3-4.8.5. Retention times for each ncBCAA were
determined (Table 24) and calibration curves were elaborated by plotting the ratio of the analyte
signal (ncBCAA) to the internal standard signal (ABA) as a function of the analyte concentration
(Figure 16).
Figure 16. Calibration curves for ncBCAA. The y-axis in the calibration curves represents the ratio of the area of
a given peak corresponding to one amino acid by the peak area of ABA (IncBCAA/IABA(200 µM)). The x-axis refers to
the real amino acid concentration. Coefficient of determination (R2) and linear regression equation are also
shown.
5.5.2 Elaboration of calibration curves and determination of retention times for
canonical amino acids
The following amino acid mixtures were prepared for calibration: calibration level 1 (50 µM): 25 μL
200 μM SD1 or SD2 plus 100 μL 200 μM ABA; calibration level 2 (100 µM): 50 μL 200 μM SD1 or SD2
plus 100 μL 200 μM ABA; calibration level 3 (200 µM): 100 μL 200 μM SD1 or SD2 plus 100 μL 200 μM
ABA; calibration level 4 (400 µM): 200 μL 200 μM SD1 or SD2 plus 100 μL 200 μM ABA. Each mixture
was used for amino acid analysis according to sections 4.8.3-4.8.5. Retention times for each canonical
amino acid were determined (Table 24) and calibration curves were elaborated by plotting the ratio
of the analyte signal (canonical amino acid) to the internal standard signal (ABA) as a function of the
analyte concentration (Figure 17).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
89
Figure 17. Calibration curves for canonical amino acids. The y-axis represents the ratio of the area of a given
peak corresponding to one amino acid by the peak area of ABA (IAA/IABA(200 µM)). The x-axis refers to the real
amino acid concentration in µM units. Coefficient of determination (R2) and linear regression equation are also
shown.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
90
Results
Table 24. Retention times for each tested amino acid.
Amino acid
Retention time (min)
Amino acid
Retention time (min)
alanine
2.461
threonine
3.447
glycine
2.601
serine
3.517
2-aminobutyric acid
2.728
proline
4.190
valine
2.839
aspartate
4.210
norvaline
3.099
methionine
4.550
leucine
3.158
glutamate
4.584
isoleucine
3.282
phenylalanine
5.843
norleucine
3.340
lysine
6.033
β-methylnorleucine
3.398 and 3.447
histidine
6.311
tyrosine
6.307
5.5.3 Evaluation of the effect of hydrolysis on amino acid analysis
For amino acid analysis of protein samples, a hydrolysis approach has to be previously followed in
order to obtain the free amino acids. The most employed hydrolysis method is the acid hydrolysis
with 6N HCl. However, according to literature (Pickering and Newton, 1990; Davidson, 1997), the use
of acid hydrolysis can have various negative impacts affecting subsequent amino acid analysis, which
needed to be evaluated in this work. In order to assess the effects of acid hydrolysis, amino acid
content of standard solutions, being or not subjected to the hydrolysis process, was analyzed by GC-
FID and results were compared. The standard solutions consisted of 200 µM ncBCAA, 200 µM SD1
and 200 µM SD2. The standard solution ncBCAA contained amino acids norvaline, norleucine and β-
methylnorleucine, the standard solution SD2 comprised amino acids asparagine, glutamine and
tryptophan and standard solution SD1 contained the remaining canonical amino acids.
For samples subjected to hydrolysis, the hydrolysis protocol was applied as described in section 4.8.2.
In the current case, 200 μL of the corresponding standard solution were mixed with 800 μL 5M HCl.
The resulting hydrolyzed sample pellets were then resuspended with 200 μL of a solution containing
20 mM HCl and 10% isopropanol. 100 μL of the resulting suspension were then mixed with 100 μL of
200 μM ABA and these solutions were used for amino acid analysis according to described in sections
4.8.3-4.8.5. For samples not subjected to hydrolysis, protocol applied was as follows: 100 μL of the
corresponding standard solution were mixed with 100 μL of 200 μM ABA and resulting solutions
were used for amino acid analysis according to described in sections 4.8.3-4.8.5. Amino acid content
from both solution sets was analyzed by GC-FID and the % of variation of the determined amino acid
concentration of hydrolyzed amino acid samples with respect to non-hydrolyzed samples was
calculated for each amino acid (Figure 18).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
91
For most of the amino acids, a signal reduction was reported after hydrolysis, being this especially
evident for norvaline, norleucine, threonine, methionine and tyrosine. Interestingly, the signal of
glutamate reported an increase after hydrolysis. In addition, asparagine and glutamine were
respectively deamidated to aspartate and glutamate after hydrolysis, while tryptophan was
destroyed.
Figure 18. Charts representing the % of variation of the determined amino acid concentration of hydrolyzed
amino acid samples with respect to non-hydrolyzed samples for each amino acid. The chart in the left presents
the % of variation of amino acids present in standard solutions SD1 and ncBCAA while the chart in the right
shows the % of variation of amino acids present in standard solution SD2.
5.5.4 Validation of the GC-FID method by analyzing a pure protein
In order to validate the established GC-FID method, a pure pre-proinsulin (PPI) variant, whose amino
acid composition and molecular weight are known, was employed for amino acid analysis by GC-FID.
Solution containing pure PPI was hydrolyzed as described in section 4.8.2. In the current case, 250 μL
of a 10 mg/mL PPI solution were mixed with 750 μL 5M HCl. Hydrolysis was performed at both 24 h
and 72 h in order to identify if hydrolysis time has an effect on amino acid analysis. The hydrolyzed
protein pellet was resuspended with 250 μL dH2O. Protein concentration of the resulting solution
was measured by Nanodrop and amino acid concentration was then calculated according to
following equation: 𝐴𝑚𝑖𝑛𝑜 𝑎𝑐𝑖𝑑 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 (µ𝑚𝑜𝑙
µ𝐿 ) = 𝑃𝑟𝑜𝑡𝑒𝑖𝑛 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 (𝑚𝑔
𝑚𝐿)∙𝑚𝑜𝑙𝑒𝑠 𝑎𝑚𝑖𝑛𝑜 𝑎𝑐𝑖𝑑
𝑃𝑟𝑜𝑡𝑒𝑖𝑛 𝑚𝑜𝑙𝑒𝑐𝑢𝑙𝑎𝑟 𝑤𝑒𝑖𝑔ℎ𝑡 (𝑘𝐷𝑎)∙103.
The volume of hydrolyzed PPI solution corresponding to 10 µmol amino acids was mixed with 100 µL
200 µM ABA. The resulting samples were treated according to described in sections 4.8.3-4.8.5 and
amino acid concentrations were determined. From that data, the experimental amino acid
composition of the PPI was calculated and this was compared to the theoretical amino acid
composition of PPI (Figure 19). Alanine was used as the reference amino acid for those calculations
since, according to results obtained at the previous section 5.5.3, alanine was the amino acid less
affected by hydrolysis.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
92
Results
Figure 19. Comparison of the theoretical amino acid composition of PPI (blue) with the experimental amino
acid composition of PPI calculated after GC-FID analysis of pure PPI samples subjected to 24h (red) or 72h
(green) acid hydrolysis.
Experimental results matched quite accurately the theoretical amino acid composition. Moreover, 72
h hydrolysis slightly improved results: leucine concentration increased up to theoretical values;
threonine, serine, proline and glutamate concentrations also increased, reaching values slightly
above theoretical values; valine, isoleucine, phenylalanine and lysine concentrations increased as
well, but they still remained below real values. However, since 72 h hydrolysis requires 2 additional
days and the improvement of hydrolysis performance is not so dramatic, 24 h hydrolysis was selected
for further experimentation.
5.6 Establishment of cultivation conditions based on pyruvate pulsing
leading to an increase of ncBCAA mis-incorporation into recombinant
mini-proinsulin expressed in E. coli at shake flask level
The aim of this experiment was to mimick large-scale effects, i.e. increase of ncBCAA biosynthesis
and minisincorporation into recombinant mini-proinsulin, at shake flask level by exposing an E. coli K-
12 BW25113 pSW3_lacI+ cultivation to pyruvate pulses of different intensity.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
93
5.6.1 Cultivation conditions
Pre-culture was prepared as follows: 20 mL of TUB medium were introduced in a sterile 100 mL
Erlenmeyer shake flask. Ampicillin was supplemented to a final concentration of 100 µg/mL. Fresh
cells from strain E. coli K-12 BW25113 pSW3_lacI+ were inoculated into the medium and culture was
incubated at 37 °C and 220 rpm, overnight.
The main culture was prepared and cultivated as follows: 7 sterile 300 mL Erlenmeyer shake flasks
were filled with 50 mL TUB medium. OD600nm of the pre-culture was measured and the volume from
the pre-culture needed to generate a main culture at 0.1 OD600nm in a final volume of 50 mL was
calculated. The corresponding calculated volume of pre-culture was introduced in the 300 mL sterile
Erlenmeyer shake flasks. Ampicillin was supplemented to a final concentration of 100 µg/mL.
Resulting cultures were incubated at 37°C and 220 rpm. When OD600nm reached 0.3, pulses of
different pyruvate concentrations were applied each 30 minutes for a time period of 2.5h (a total of
6 pyruvate pulses). Seven different pyruvate concentrations were tested, one for each available
shake flask: 0, 100, 250, 500, 1000, 2000 and 5000 mg/L. Induction with 0.5 mM IPTG was performed
when cultures reached OD600nm of 0.6. Around 3.5h after induction cultivations were stopped, OD600nm
was measured and cells contained in the whole liquid culture were harvested by centrifugation at
10000 g for 10 min for further amino acid analysis by GC-FID.
5.6.2 Analysis of ncBCAA
Intracellular soluble protein fraction and inclusion body isolation from cell extracts was carried out
according to described in section 4.8.1. Acid hydrolysis of intracellular soluble and inclusion body
protein fractions was performed as described in section 4.8.2. In the current case, 250 μL of the
intracellular soluble protein fraction were mixed with 750 μL 5M HCl. Isolated inclusion body pellets
were resuspended with 200 μL H2O and afterwards 750 μL 5M HCl were added. Hydrolyzed IB pellet
samples were resuspended with 500 μL dH2O while hydrolyzed intracellular soluble protein fraction
samples were resuspended with 1 mL dH2O. The resulting hydrolyzed samples were treated
according to described in section 4.8.3. In the current case, 50 µL of the hydrolyzed intracellular
soluble protein fraction or the hydrolyzed inclusion body fraction were mixed with 100 µL of 200 µM
ABA. The resulting samples were used for amino acid analysis by GC-FID according to the procedure
described in sections 4.8.4-4.8.5. Concentrations of ncBCAA in both inclusion body and intracellular
soluble fractions under the different tested cultivation conditions are shown in Figure 20.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
94
Results
Figure 20. Molar concentrations of norleucine (red bars), norvaline (blue bars) and β-methylnorleucine (green
bars) normalized to OD600nm present in the intracellular soluble protein fraction (A) and inclusion body fraction
(B) from samples taken 3.5 h after IPTG induction of E. coli K-12 BW25113 pSW3_lacI+ cultivations in shake
flasks under cultivation conditions subjected to pulses of different pyruvate concentration. When cultivation
achieved OD600nm of 0.3, pyruvate was supplemented in a pulse-based manner each 30 minutes for a time
period of 2.5h, this is a total of 6 pyruvate pulses. Depending on the cultivation pyruvate pulses were added up
to 0, 100, 250, 500, 1000, 2000 and 5000 mg/L. Induction with 0.5 mM IPTG was performed when cultures
reached OD600nm of 0.6.
For both intracellular soluble protein and inclusion body fractions, addition of increasing
concentrations of pyruvate triggered norleucine formation. However, it is noteworthy to point out
that norleucine concentration calculated in the intracellular soluble fraction under 5 g/L pyruvate
was clearly out of the general trend. This value was considered as an outlier and not employed for
further data interpretation. The positive effect of pyruvate in norvaline formation was less evident. In
the inclusion body fraction, a positive trend was shown from 250 mg/L to 5 g/L pyruvate. However, in
the intracellular soluble fraction a clear trend could not observed due to the high variance. This
observation also applied for β-methylnorleucine (Figure 20). Similar results were obtained when
expressing data as ratio of the concentration of ncBCAA with respect to the canonical counterpart
(Figure S70).
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
95
5.7 Establishment of cultivation conditions based on pyruvate pulsing and O2
limitation leading to an increase of ncBCAA mis-incorporation into
recombinant mini-proinsulin expressed in E. coli in a 10 mL mini-reactor
and in a 15L reactor
In section 5.6 it was proven that addition of pyruvate pulses effectively triggers an increase in
norleucine biosynthesis and mis-incorporation when cultivating E. coli K-12 BW25113 expressing
plasmid pSW3_lacI+ at shake flask level. However, such effect was not clearly shown for the ncBCAA
norvaline and β-methylnorleucine. It might be that cultivation conditions present at shake flask level
were not optimal to clearly evaluate the effect of pyruvate on ncBCAA formation. Hence, the next
step in this investigation was to perform similar experiments with pyruvate in more controlled
cultivation systems such as mini-reactor and 15L reactor.
5.7.1 Experiment in a 10 mL mini-reactor
5.7.1.1 Cultivation mode
5.7.1.1.1 Standard cultivation
30 µL of a cryostock containing E. coli K-12 BW25113 pSW3_lacI+ were used to inoculate 30 mL of 1:3
supplemented TUB mineral salt medium containing 5 g/L glucose, 0.1 M Na-Phosphate buffer and
100 µg/mL ampicillin in order to generate the pre-culture. The pre-culture was incubated at 37 °C
and 220 rpm in an orbital shaker, overnight. OD600nm at the end of the pre-culture was measured and
a given volume was used to inoculate a 5 mL-starting volume mini-reactor so that initial OD600nm was
0.4. The mini-reactor medium consisted of a 1:3 supplemented TUB mineral salt medium containing
4 g/L glucose, 0.1 M Na-Phosphate buffer, 100 µg/mL ampicillin and 1 µL/mL antifoam Desmophen.
Cultivation was carried out at 37 °C and the pH was maintained at 7 by automatic control with
NH4OH and CO2. Stirrer speed was set to 800 rpm and DO set-point to 25 %, maintaining the last by
automatically increasing the oxygen flow into the mini-reactor. Batch phase lasted around 4h. After
batch phase was finished, 1 mL 400 g/L EnPump 200 solution and 50 µL 3000 U/L amylase solution
were manually added into the mini-reactor, hence starting the fed-batch phase. EnPump 200 is a
glucose polymer and when amylase is present, it constantly hydrolyses the polymer, thus delivering
free glucose molecules over time, ensuring then a glucose-limited fermentation. In order to generate
the 400 g/L EnPump 200 solution, 25 g EnPump 200 powder was dissolved in a 25 mL solution, having
the same composition than the medium already present in the mini-reactor, so that components’
concentration in the mini-reactor remained invariable after adding the EnPump 200 solution. 30 min
after beginning of the fed batch phase, expression of recombinant mini-proinsulin was induced by
manual addition of an IPTG pulse to a final concentration of 0.5 mM. Fed-batch phase was active for
3.5 h. A general overview of the cultivation is shown in Figure 21. This cultivation was performed in
triplicates in mini-reactor wells A1, A2 and A3.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
96
Results
Figure 21. Overview of the standard cultivation of E. coli K-12 BW25113 pSW3_lacI+ in a 10mL mini-reactor. A 4
h batch phase was followed by a 3.5 h EnPump-based fed- batch phase. Induction was carried out 30 min after
fed-batch began (at 5h cultivation time). Cultivation was carried out at 37 °C and the pH was maintained at 7 by
automatic control with NH4OH and CO2. Stirrer speed was set to 800 rpm and DO set-point to 25 %, controlling
the last by automatically modifying the oxygen flow into the mini-reactor.
5.7.1.1.2 Cultivation triggering increase of ncBCAA production
Cultivation was performed as described for the standard cultivation in section 5.7.1.1.1. However,
immediately after beginning of the fed-batch phase, a 0.833 g/L pyruvate pulse was manually added
into the reactor. During the following 5 min after pyruvate addition, DO set-point was set to 0, so
that no oxygen was supplied into the mini-reactor during that period, hence ensuring oxygen
limitation. 30 min after the first pyruvate pulse, expression of recombinant mini-proinsulin was
induced by manual addition of an IPTG pulse to a final concentration of 0.5 mM. After induction,
sequential 0.833 g/L pyruvate pulses were manually performed each 30 min as described above for a
total of 5 pulses. Between pulses, DO set-point was re-established to 25 %. Fed-batch phase was
active for 3.5 h. A general overview of the cultivation is shown in Figure 22. This cultivation was
performed in triplicates in mini-reactor wells A4, A5 and A6.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
97
Figure 22. Overview of the cultivation of E. coli K-12 BW25113 pSW3_lacI+ in a 10mL mini-reactor under
conditions triggering increase of ncBCAA production, during the whole cultivation process (A) and during the
fed-batch period (B). A 4 h batch phase was followed by a 3.5 h EnPump-based fed-batch phase. Induction was
carried out 30 min after fed-batch began (at 5h cultivation time). Right after beginning of fed-batch phase and
then each 30 min, a pyruvate pulse was added (indicated by orange arrows in B). 6 pyruvate pulses
corresponding to 5 g/L pyruvate were added in total. During the following 5 min after pyruvate addition, DO
set-point was set to 0, so that no oxygen was supplied into the mini-reactor. Cultivation was carried out at 37
°C and the pH was maintained at 7 by automatic control with NH4OH and CO2. Stirrer speed was set to 800 rpm
and DO set-point to 25 %, controlling the last by automatically modifying the oxygen flow into the mini-reactor.
A
B
B
B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
98
Results
5.7.1.2 Mini-proinsulin analysis by SDS-PAGE from total cell extracts
Recombinant mini-proinsulin expression was evaluated by SDS-PAGE by densitometry analysis.
Samples consisted on OD1/mL cell pellets and were taken 3h after induction. Samples were
processed for SDS-PAGE analysis according to described in section 4.7.15. Scanned 2-dimensional
(2D) SDS-PAGE gels are shown in Figure 23, where each gel lane corresponds to a cell suspension of
OD600nm=0.25. In order to allow quantification of mini-proinsulin content in the samples, a pure pre-
proinsulin (PPI) calibrate was employed. Hence, 1 µg, 0.5 µg and 0.25 µg pure pre-proinsulin were
loaded into each gel.
Figure 23. 2D SDS-PAGE gels of total cell extracts from samples A1, A2, A3, A4, A5 (A) and A6 (B). Each gel lane
corresponds to a cell suspension of OD600nm=0.25. In order to allow quantification of recombinant mini-
proinsulin content in the samples by densitometry analysis, a pure pre-proinsulin having the same molecular
weight than the expressed recombinant protein was employed. Hence, 1 µg, 0.5 µg and 0.25 µg pure pre-
proinsulin were loaded into each gel. Std refers to the protein standard SeeBlue® Plus2 Pre-stained Protein
Standard, shown in kDa units. Samples A1, A2 and A3 are triplicates of the standard cultivation while A4, A5
and A6 are triplicates of the cultivation triggering increase of ncBCAA production.
A 3 point calibration curve was then elaborated for each gel by correlating added pre-proinsulin mass
(1 µg, 0.5 µg and 0.25 µg pure PPI) with its corresponding band intensity in the SDS-PAGE gel (Figure
24).
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
99
Figure 24. Calibration curve correlating added pre-proinsulin mass and its corresponding band intensity in the
SDS-PAGE gel containing samples A1, A2, A3, A4, A5 (left) and A6 (right). Samples A1, A2 and A3 are triplicates
of the standard cultivation while A4, A5 and A6 are triplicates of the cultivation triggering increase of ncBCAA
production.
Afterwards, the intensity of the band corresponding to the expressed recombinant mini-proinsulin in
a given sample was interpolated into the previously generated calibration curves and the
corresponding PPI mass was then estimated. PPI concentrations resulting from the densitometry
analysis are shown in Figure 25. The strategy based on pyruvate pulsing and O2 limitation did not
have a negative impact on recombinant protein production, as seen by the average values.
Figure 25. Estimated concentrations of recombinant mini-proinsulin 3h after induction of E. coli K-12 BW25113
pSW3_lacI+ cultivations from mini-reactor wells A1-A6 (A) and its average (B). Samples A1, A2 and A3 are
triplicates of the standard cultivation (-) while A4, A5 and A6 are triplicates of the cultivation triggering increase
of ncBCAA production (+).
5.7.1.3 Mini-proinsulin analysis by SDS-PAGE from inclusion body fractions
Culture broths from mini-reactor wells A1 to A6 were treated as previously described in section 4.8.1
in order to isolate inclusion bodies containing recombinant mini-proinsulin. In order to evaluate
efficiency of the IB isolation process, samples after inclusion body isolation were analyzed by SDS-
A B
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
100
Results
PAGE as described in step 12 in section 4.8.1. Scanned 2D SDS-PAGE gels are shown in Figure 26A,
where each gel lane corresponds to a cell suspension of OD600nm=0.5. As a control, the total cell
extract of an OD600nm=0.5 cell suspension from mini-reactor A1 was employed.
By comparing cell extract and inclusion body samples it could be concluded that through the
inclusion body isolation process there was no significant loss of recombinant mini-proinsulin, since
the corresponding PPI band remained quite similar, and that recombinant mini-proinsulin in the
inclusion body fraction was quite pure. Additionally, a 3-dimensional (3D) image showing volumetric
intensities of bands present in the gel was also taken in order to clearly see that the recombinant
mini-proinsulin was the predominant protein in the inclusion body fraction (Figure 26B).
Figure 26. SDS-PAGE gel of inclusion body fraction from samples A1 to A6 in 2D (A) and 3D (B). Each gel lane
corresponds to a cell suspension of OD600nm=0.5. In order to determine efficiency of the inclusion body isolation
process a control consisting on total cell extract of an OD600nm=0.5 cell suspension from mini-reactor A1 was
included. Std refers to the protein standard SeeBlue® Plus2 Pre-stained Protein Standard, shown in kDa units.
Samples A1, A2 and A3 are triplicates of the standard cultivation while A4, A5 and A6 are triplicates of the
cultivation triggering increase of ncBCAA production.
Figure 27. Chart showing band intensity (y-axis) over lane position (x-axis) for SDS-PAGE lanes “A1 cell extract”
(A) and “A1 IB” (B). Lane region containing expressed recombinant mini-proinsulin was selected (highlighted in
green) and the percentage of lane intensity represented by the recombinant protein was estimated, obtaining
25.4 % for (A) and 80.9% for (B).
For lanes “A1 cell extract” and “A1 IB” a chart showing intensity over lane position was also
generated in order to estimate the percentage of the lane intensity represented by the recombinant
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
101
mini-proinsulin (Figure 27). In the cell extract, mini-proinsulin represented about 25% of the total
protein while, after inclusion body isolation, this percentage increased up to about 81%. After results
shown in this section, it was concluded that efficiency of performed inclusion body isolation process
was high, obtaining a quite pure recombinant protein for further amino acid analysis.
5.7.1.4 Analysis of ncBCAA
Intracellular soluble protein fraction and inclusion body isolation from total cell extracts was carried
out according to described in section 4.8.1. In the current case, at step 1, OD600nm at the end of the
cultivation was determined and all samples were standardized so that OD600nm was equal to 43.7 in 1
mL. The corresponding volume was calculated for each sample and cells were harvested from liquid
culture by centrifugation at 10000 g for 10 min. Acid hydrolysis of intracellular soluble and inclusion
body protein fractions was performed as described in section 4.8.2. In the current case, 250 μL of the
intracellular soluble protein fraction were mixed with 750 μL 5M HCl. Isolated inclusion body pellets
were resuspended with 200 μL H2O. 100 μL of the resulting inclusion body suspension were mixed
with 900 μL 5M HCl. Hydrolysed IB pellet samples were resuspended with 500 µL dH2O while
hydrolysed intracellular soluble protein fraction samples were resuspended with 1 mL dH2O. The
resulting hydrolyzed samples were treated according to described in section 4.8.3. In the current
case, 200 μL of the resulting IB solution were mixed with 100 μL 200 µM ABA while 100 μL of the
resulting solution containing intracellular soluble protein fraction were mixed with 100 μL 200 µM
ABA. The resulting samples were used for amino acid analysis by GC-FID according to the procedure
described in sections 4.8.4-4.8.5. Concentrations of ncBCAA in both inclusion body and intracellular
protein soluble fractions under the 2 tested cultivation conditions are shown in Figure 28.
Figure 28. Molar concentrations of norvaline (blue bars), norleucine (red bars) and β-methylnorleucine (green
bars) normalized to OD600nm present in the intracellular soluble protein fraction (A) and inclusion body fraction
(B) at 3 h after induction of E. coli K-12 BW25113 pSW3_lacI+ cultivation in a 10 mL mini-reactor under standard
cultivation conditions (-) or under conditions triggering ncBCAA production, i.e. pyruvate pulses combined with
O2 limitation (+).
After implementing pyruvate pulses combined with O2 limitation at mini-reactor level, norvaline
increased a 17.5 %, norleucine increased a 51.7 % and β-methylnorleucine decreased a 14.8 % in the
intracellular soluble protein fraction while norvaline increased a 25.6 % and norleucine increased a
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
102
Results
28.1 % in the inclusion body fraction when normalizing amino acid concentration to OD600nm (Figure
28). After implementing pyruvate pulses combined with O2 limitation at mini-reactor level, norvaline
increased a 73.7 % while norleucine increased a 32.9 % in the inclusion body fraction when
normalizing amino acid concentration to PPI mass (Figure S71). Similar results were obtained when
expressing data as ratio of the concentration of ncBCAA with respect to the canonical counterpart
(Figure S72).
5.7.2 Experiment in a 15 L reactor
5.7.2.1 Cultivation mode
5.7.2.1.1 Standard cultivation
100 µL of a cryostock containing E. coli K-12 BW25113 pSW3_lacI+ were used to inoculate 500 mL of
supplemented TUB mineral salt medium containing 5 g/L glucose and 100 µg/mL ampicillin in order
to generate the pre-culture. The pre-culture was incubated at 37 °C and 220 rpm in an orbital shaker
for 12 h, using an initial cold-start technique. OD600nm at the end of the pre-culture was measured and
a given volume was used to inoculate a 7 L starting volume reactor so that initial OD600nm was 0.4. The
reactor medium consisted of supplemented TUB mineral salt medium containing 5 g/L glucose, 2 mL
antifoam (Antifoam 2014, Sigma) and 100 µg/mL ampicillin. Cultivation was carried out at 37 °C and
the pH was maintained at 7 by automatic control with 25% NH4OH. Airflow was set to 7 vvm and DO
set-point to 20 %, maintaining the last by using a cascade control altering stirrer speed (initial stirrer
speed was set to 800 rpm). Batch phase lasted effectively 4h, with an intermediate 13h cold phase at
15 °C. At the end of the batch phase, exponential feeding was started, according to following
equation,
𝐹(𝑡)=𝑞𝑠
𝑆∙(𝑋∙𝑉)∙𝑒µ𝑠𝑒𝑡 ∙ 𝑡
where
F (t)
represents the feed flow rate over time (L h-1),
qs
the set-point of the specific substrate
uptake rate (0.514 gS gX-1 h-1),
S
the concentration of glucose in the feed solution (442 g/L),
X
the
biomass concentration over time (g/L),
V
the volume of the reactor over time (L), µ
set
the set-point
of the specific cell growth rate (0.3 h-1) and
t
the time during the fed-batch phase. The feed solution
consisted of TUB mineral salt medium supplemented with 4 mL/L trace elements solution, 2 mL/L
MgSO4 solution (1.0 M), 100 µg/mL ampicillin and 442 g/L glucose.
Exponential fed-batch phase was carried out for 3 hours and afterwards expression of recombinant
mini-proinsulin was induced by automatic addition of IPTG to a final concentration of 0.5 mM.
Induction time was 30 minutes. During the induction phase no feed was added into the reactor. After
induction, a constant feeding phase was started, so that the constant flow rate was equal to the last
flow rate achieved in the exponential feeding phase. Constant feed fed-batch phase was active for 5-
6 h. Glucose concentration in the reactor was determined by using a BioPAT® Trace online glucose
analyser. A general overview of the cultivation is shown in Figure 29.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
103
Figure 29. Overview of the standard cultivation of E. coli K-12 BW25113 pSW3_lacI+ in a 15L reactor, during the
whole cultivation process (A) and during fed-batch period (B). Different cultivation phases are shown in the
diagram as BE (first 2h of batch phase), KUL (13h cold period at 15 °C), FE (remaining 2h of batch phase), F1 (3h
exponential fed-batch phase) and IND (induction and linear fed-batch phase). IPTG induction was performed
during 30 minutes (20.75 to 21.25 h cultivation time). Present in the diagram axes, Flow1 corresponds to the
flow rate (g/h) of the pump transporting the feed solution into the reactor while Flow2Y corresponds to one
tenth of the flow rate (g/h) of the pump transporting the IPTG solution used for induction. During the linear
feeding phase the constant flow rate was supposed to be equal to the last flow rate achieved in the
exponential feeding phase; however, during the beginning of the linear fed-batch phase that flow rate was
higher due to a programming error. That explains the sudden accumulation of glucose (up to 0.8 g/L) between
22 and 23h cultivation time. Nevertheless, error was corrected by setting the flow rate to the proper value and
glucose limitation was then again rapidly restored.
A
B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
104
Results
5.7.2.1.2 Cultivation triggering increase of ncBCAA production
Cultivation was performed as described for the standard cultivation in section 5.7.2.1.1. However, in
the current cultivation the feed solution contained 454 g/L glucose instead of 442 g/L. Moreover,
after the exponential fed-batch phase, 1 g/L pyruvate pulse was automatically added into the
reactor. Pyruvate solution was constantly pumped for 5 minutes. During that time period no feed
was added, airflow rate was temporary set to 0 and DO cascade control was disconnected. After the
first pyruvate pulse, expression of recombinant mini-proinsulin was induced by automatic addition of
IPTG to a final concentration of 0.5 mM. Induction time was 30 minutes. During the induction phase
no feed was added into the reactor and airflow and DO cascade control were re-established. After
induction, sequential 1 g/L pyruvate pulses were performed each 30 min as described above for a
total of 4 pulses. Between pulses, constant feeding phase was activated, so that the constant flow
rate was equal to the last flow rate achieved in the exponential feeding phase, and airflow and DO
cascade control were re-established. Constant feed fed-batch phase was active for 5-6 h. A general
overview of the cultivation is shown in Figure 30.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
105
Figure 30. Overview of the cultivation of E. coli K-12 BW25113 pSW3_lacI+ in a 15L reactor under conditions
triggering increase of ncBCAA production, during the whole cultivation process (A), during fed-batch period (B)
A
B
C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
106
Results
and during pyruvate pulsing (C). Different cultivation phases are shown in the diagram as BE (first 2h of batch
phase), KUL (10h cold period at 15 °C), FE (remaining 2h of batch phase), F1 (3h exponential fed-batch phase)
and IND (induction, linear fed-batch phase and pyruvate pulsing). IPTG induction was performed for 30 minutes
(18.25 to 18.75 h cultivation time). Pyruvate pulses are indicated by orange arrows (C). Present in the diagram
axes, Flow1 corresponds to the flow rate (g/h) of the pump transporting the feed solution into the reactor
while Flow2Y corresponds to one tenth of the flow rate (g/h) of the pump transporting either the IPTG solution
used for induction or the pyruvate solution employed for pulsing. Exponential fed-batch was, due to a
programming error, started shortly before glucose was completely consumed in the batch phase. Hence a small
glucose accumulation was reported at around 15h. Thus, exponential feeding was shortly shut down until
glucose was completely depleted and then, activated again.
5.7.2.2 Mini-proinsulin analysis by HPLC
Concentration of recombinant mini-proinsulin from hourly samples was analyzed according to an in
house HPLC method. 1 mL of culture broth was directly used for analysis. PPI concentrations resulting
from the in house HPLC analytical method are shown in Figure 31. PPI concentrations determined
under both tested cultivation conditions were really similar. In addition, cell growth behavior
(represented by OD600nm and CDW) was also comparable for both tested cultivation conditions (Figure
S73), suggesting that specific production of recombinant mini-proinsulin remained the same,
independently of the cultivation strategy employed.
Figure 31. Estimated concentrations of recombinant mini-proinsulin over time after induction of E. coli K-12
BW25113 pSW3_lacI+ cultivation (WT E. coli) in a 15L reactor under standard conditions (STD) and under
conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-O2).
5.7.2.3 Analysis of ncBCAA
Intracellular soluble protein fraction and inclusion body isolation from total cell extracts was carried
out according to described in section 4.8.1. In the current case, at step 1, the OD600nm of the culture
at a given time point was determined and all samples were standardized so that OD600nm was equal to
50 in 1 mL. The corresponding volume was calculated for each sample and cells were harvested from
liquid culture by centrifugation at 10000 g for 10 min. Acid hydrolysis of intracellular soluble protein
and inclusion body fractions was performed as described in section 4.8.2. In the current case, 250 μL
of the intracellular soluble protein fraction were mixed with 750 μL 5M HCl. Isolated inclusion body
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
107
pellets were resuspended with 200 μL H2O. 100 μL of the resulting inclusion body suspension were
mixed with 900 μL 5M HCl. After acid hydrolysis and speed-back, hydrolysed IB pellet samples were
resuspended with 250 μL dH2O while hydrolysed intracellular soluble protein fraction samples were
resuspended with 1 mL dH2O. The resulting hydrolyzed samples were treated according to described
in section 4.8.3. In the current case, 250 μL of the resulting IB solution (whole solution) were mixed
with 100 μL 200 µM ABA while 100 μL of the resulting solution containing intracellular soluble
protein fraction were mixed with 100 μL 200 µM ABA. The resulting samples were used for amino
acid analysis by GC-FID according to the procedure described in sections 4.8.4-4.8.5. Concentrations
of ncBCAA in the intracellular soluble protein and inclusion body fraction over cultivation time are
shown in Figure 32 and Figure 33, respectively.
Figure 32. Molar concentrations of norvaline (A), norleucine (B) and β-methylnorleucine (C) normalized to
OD600nm present in the intracellular soluble protein fraction calculated over time after induction of E. coli K-12
BW25113 pSW3_lacI+ cultivation (WT E. coli) in a 15L reactor under standard conditions (STD) and under
conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-O2). Orange
arrows indicate time points where 1 g/L pyruvate pulse combined with 5 min O2 limitation was applied.
A B
C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
108
Results
For both tested cultivations, a progressive accumulation of norleucine and β-methylnorleucine was
reported in the intracellular soluble protein fraction over time after induction, being it more
significant for norleucine. However, for standard cultivation “WT E.coli, STD”, after an initial increase
reported at 0 h induction time, norvaline concentration remained quite invariable over time (Figure
32).
Cultivation subjected to pyruvate pulses and O2 limitation (“WT E.coli, PYR-O2”) translated in an
increase of ncBCAA with respect to cultivation “WT E.coli, STD”. Depending on the analyzed ncBCAA,
such increase was reported at different time points after first pyruvate pulse was applied. Norvaline
and norleucine concentrations increased in cultivation “WT E.coli, PYR-O2” with respect to “WT
E.coli, STD” 1.5 h after first pyruvate pulse was applied (1 h after induction) and such increase
persisted for the remaining cultivation time. However, increase of β-methylnorleucine concentration
was reported from 4.5 h after first pyruvate pulse (4 h after induction) on. This suggests that
pyruvate is not immediately metabolized after its addition into the reactor and that the metabolic
effect triggered by pyruvate and O2 limitation has a limited influence on β-methylnorleucine
synthesis, in contrast to norleucine and norvaline (Figure 32).
Furthermore, norvaline concentration increased progressively in cultivation “WT E.coli, PYR-O2” until
3h after induction. However, from that time point on, norvaline concentration progressively dropped
to finally reach values comparable to standard cultivation “WT E.coli, STD” 5 h after induction. Hence,
the norvaline increase resulting upon applying pyruvate pulses and O2 limitation disappeared 2h after
last pyruvate pulse was applied. Interestingly, this observation was not shown for norleucine and β-
methylnorleucine, which reported a progressive accumulation over cultivation time (Figure 32).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
109
Figure 33. Molar concentrations of norvaline (A, C) and norleucine (B, D) normalized to OD600nm (A, B) or to PPI
mass (C, D) present in the inclusion body fraction calculated over time after induction of E. coli K-12 BW25113
pSW3_lacI+ cultivation (WT E. coli) in a 15L reactor under standard conditions (STD) and under conditions
triggering ncBCAA accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-O2). Orange arrows indicate
time points where 1 g/L pyruvate pulse combined with 5 min O2 limitation was applied.
For both tested cultivations, a progressive accumulation of norvaline and norleucine was reported in
the inclusion body fraction over time after induction. Until 2 h after induction, norvaline and
norleucine concentrations remained similar in both tested cultivations. However, from that time
point on, norvaline and norleucine concentrations analyzed in cultivation “WT E.coli, PYR-O2”
remained higher than in “WT E.coli, STD” (Figure 33). Hence, the increase of norvaline and norleucine
concentration resulting upon applying pyruvate pulses and O2 limitation was first reported 3.5 h after
first pyruvate pulse was applied (3 h after induction), which is 2 h later than reported in the
intracellular soluble protein fraction. This suggests that metabolic effects triggered by pyruvate
pulsing combined with O2 limitation occur first in the cytosol (intracellular soluble protein fraction),
where the ncBCAA are synthesized. Afterwards, translation machinery would mis-incorporate those
ncBCAA present in the cytosol into the nascent recombinant proteins (inclusion body fraction). β-
C D
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
110
Results
methylnorleucine could not be properly detected in the inclusion body fraction since at the expected
retention time in the chromatogram no peak was found (Figure 33).
5.7.2.4 Acetate and formate analysis
Acetate and formate concentrations from hourly samples were analysed offline by in house
enzymatic assays available at Sanofi-Aventis Deutschland GmbH. Sample preparation for acetate and
formate analysis was as follows: 10 mL of culture broth were inactivated with Bardac, centrifuged at
4000 g for 10 min and resultant supernatant was decanted and filtered through a 0.45 µm filter. The
filtrate was then used for analysis. Estimated concentrations of acetate and formate in culture broth
supernatants over cultivation time are shown in Figure 34.
Figure 34. Concentration of acetate (A) and formate (B) present in culture broth supernatants calculated over
fed-batch time of cultivation E. coli K-12 BW25113 pSW3_lacI+ (WT E. coli) in a 15L reactor under standard
cultivation conditions (STD) and under conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and
oxygen limitation (PYR-O2). Orange arrows indicate time points where 1 g/L pyruvate pulse combined with 5
min O2 limitation was applied. Green arrow points out time point where IPTG induction was carried out.
B
A
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
111
For both tested cultivations, acetate concentration reached a maximum after batch phase was
finished (0 h fed-batch time). For standard cultivation “WT E.coli, STD”, acetate concentration
strongly decreased during the first hours of the fed-batch phase, reaching the minimum at around 4
h fed-batch time, to suddenly increase again, reporting a maximum at 5.5 h fed-batch time.
Afterwards, acetate decreased again and remained invariable at values around 0.2 g/L until the end
of the fed-batch phase. That sudden and momentary accumulation of acetate during the fed-batch
phase in cultivation “WT E.coli, STD” corresponds with the sudden glucose accumulation reported at
the beginning of the linear fed-batch phase due to a programming error (Figure 29). Until 4 h fed-
batch time, acetate concentration in cultivation “WT E.coli, PYR-O2” remained quite similar to “WT
E.coli, STD”. However, from that time point on, acetate concentration was always higher for
cultivation “WT E.coli, PYR-O2”, as a consequence of the metabolic alteration triggered by pyruvate
pulsing combined with O2 limitation.
For both cultivations tested, formate concentration reached a maximum after batch phase was
finished (around 0 h fed-batch time). During the first hours of fed-batch phase in standard cultivation
“WT E.coli, STD”, formate concentration decreased, but then a sudden increase was reported at 3.3
and 4.3 h fed-batch time. Afterwards, formate concentration decreased again, remaining below 0.1
g/L. Like stated for acetate, such momentary accumulation of formate in cultivation “WT E.coli, STD”
is in accordance with the sudden glucose accumulation reported at the beginning of the linear fed-
batch phase due to a programming error (Figure 29). Until 4 h fed-batch time, formate concentration
in cultivation “WT E.coli, PYR-O2” remained quite similar to “WT E.coli, STD”. However, from that
time point on, formate concentration was always higher for cultivation “WT E.coli, PYR-O2”, as a
consequence of the metabolic effect triggered by pyruvate pulsing combined with O2 limitation. Such
effect seems to be stronger for formate than acetate since, after pyruvate pulsing, increase of
formate concentration in cultivation “WT E.coli, PYR-O2” with respect to standard cultivation “WT
E.coli, STD” is about 700 %, while for acetate is about 250 %. It is also noteworthy to highlight that
right after last pyruvate pulse combined with O2 limitation was applied, acetate and formate
concentrations suddenly stopped increasing, which might suggest that anaerobic metabolism is not
active anymore.
Results derived from acetate and formate analysis are consistent with results shown in section
5.7.2.3 for norleucine and norvaline in the intracellular soluble protein fraction since effective
increase of ncBCAA in cultivation “WT E.coli, PYR-O2” with respect to standard “WT E.coli, STD”, as
reported with acetate and formate, took place 1.5 h after addition of the first pyruvate. Hence,
acetate and formate analysis were proven to be a good additional marker of ncBCAA production in E.
coli cultivations.
5.7.3 Comparison 10 mL mini-reactor and 15 L reactor
In order to elucidate if scale has an impact on ncBCAA production, a comparative study summarizing
results obtained in both scales (10 mL mini-reactor and 15 L reactor) was elaborated. Results
obtained from samples taken 3 h after induction were compared, since that time point was the only
selected for ncBCAA analysis at mini-reactor scale. In both intracellular soluble protein fraction and
inclusion body fraction, ncBCAA concentrations obtained at mini-reactor and 15 L reactor scale under
standard cultivation conditions or under conditions triggering ncBCAA production (i.e. pyruvate
pulsing combined with O2 limitation) were compared (Figure 35A and Figure 36A). In addition, for
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
112
Results
both tested scales, percentage of variation of ncBCAA concentration in cultivation triggering ncBCAA
production with respect to standard cultivation was determined (Figure 35B and Figure 36B).
Figure 35. (A) Comparison of ncBCAA concentration calculted 3h after induction in the inclusion body fraction
of E. coli K-12 BW25113 pSW3_lacI+ (WT E. coli) cultivation in 10 mL mini-reactor- and 15L reactor scale under
standard cultivation conditions (STD) or under conditions triggering ncBCAA production, i.e. pyruvate pulsing
combined with O2 limitation (PYR-O2). (B) Percentage of variation of ncBCAA concentration reported in the
cultivation triggering ncBCAA production with respect to standard cultivation.
In the inclusion body fraction, for both norvaline and norleucine, ncBCAA concentration values were
higher at 15 L reactor than at mini-reactor scale (Figure 35A). That might be explained since at bigger
scales inhomogeneities can generate different O2 and glucose gradients, hence triggering formation
of ncBCAA, even under standard cultivation conditions. In addition, effect of cultivation conditions
triggering ncBCAA production based on pyruvate pulsing combined with O2 limitation was reported
to be different depending on the tested scale, especially for norvaline (Figure 35B): concentration of
norvaline reported an increase of 26 % at mini-reactor scale while an 80 % boost was shown at 15 L
reactor scale. Norleucine concentration increased 28 % at mini-reactor scale but 21 % at 15 L reactor
scale.
Figure 36. (A) Comparison of ncBCAA concentration determined 3h after induction in the intracellular soluble
protein fraction of E. coli K-12 BW25113 pSW3_lacI+ (WT E. coli) cultivation in 10mL mini-reactor- and 15L
reactor scale under standard cultivation conditions (STD) or under conditions triggering ncBCAA production, i.e.
pyruvate pulsing combined with O2 limitation (PYR-O2). (B) Percentage of variation of ncBCAA concentration
reported in the cultivation triggering ncBCAA production with respect to standard cultivation.
A B
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
113
As reported above, in the intracellular soluble protein fraction, for both norvaline and norleucine,
concentration values were higher at 15 L reactor than at mini-reactor scale as well (Figure 36A). Also,
effect of cultivation conditions based on pyruvate pulsing combined with O2 limitation was reported
to dramatically vary depending on the tested scale (Figure 36B): concentration of norvaline reported
an increase of 17 % at mini-reactor scale while a 51 % boost was shown at 15 L reactor scale.
Norleucine concentration increased 52 % at mini-reactor scale but only 22 % at 15 L reactor scale.
However, for β-methylnorleucine effect was not so significantly different between scales when
compared with the other ncBCAA: β-methylnorleucine concentration decreased 14 % at mini-reactor
scale and reported no variation at 15 L reactor scale.
5.8 Screening of geneX-tunable E. coli strains a mini-reactor system
The aim of this experiment was to regulate the expression of strains E. coli K-12 BW25113 ΔgeneX
expressing pSW3_lacI+ and pACG_araBAD_geneX (simplified, geneX-tunable E.coli strains) by using
different concentrations of L-arabinose in a 10 mL mini-reactor. Tunable E. coli strains were grown in
Enbase-based fed-batch modus while carrying out pyruvate pulsing followed by O2 limitation,
conditions which in section 5.7 demonstrated to trigger production of ncBCAA. Tunable E. coli strains
were also tested under standard cultivation conditions. By comparing the ncBCAA production profile
obtained for each of the tested tunable E. coli strains with the control non-engineered E. coli K-12
BW25113 pSW3_lacI+ strain, it could be elucidated if the corresponding tunable E. coli strain (and
more specifically, which level of expression of the corresponding target gene) improves the product
quality of the expressed recombinant mini-proinsulin.
5.8.1 Cultivation conditions
Pre-cultures were prepared as follows: 30 µL of a cryostock containing E. coli K-12 BW25113
pSW3_lacI+ or one of the geneX-tunable E.coli strains were used to inoculate 30 mL of 1:3
supplemented TUB medium containing 5 g/L glucose, 0.1 M Na-Phosphate buffer and 100 µg/mL
ampicillin. For the tunable E.coli strains, medium also contained 25 µg/mL chloramphenicol and the
minimum L-arabinose concentration necessary to recover cell growth levels of the non-engineered E.
coli K-12 BW25113 pSW3_lacI+ strain, previously tested in section 5.3.2 (summarized in
Table 25). Pre-cultures were incubated at 37 °C and 220 rpm, overnight.
Table 25. Minimal L-arabinose concentration necessary for the cultivation of each generated tunable E.coli
strains allowing recovery of cell growth levels of the non-engineered E. coli K-12 BW25113 pSW3_lacI+ strain.
Strain
L-arabinose
concentration (%)
E. coli K-12 BW25113 pSW3_lacI+
0
E. coli K-12 BW25113 ΔleuA pSW3_lacI+ pACG_araBAD_leuA (leuA-tunable E. coli)
0.1
E. coli K-12 BW25113 ΔilvC pSW3_lacI+ pACG_araBAD_ilvC (ilvC-tunable E. coli)
0.4
E. coli K-12 BW25113 ΔthrA pSW3_lacI+ pACG_araBAD_thrA (thrA-tunable E. coli)
0.4
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
114
Results
E. coli K-12 BW25113 ΔilvA pSW3_lacI+ pACG_araBAD_ilvA (ilvA-tunable E. coli)
0,05
E. coli K-12 BW25113 ΔilvBN pSW3_lacI+ pACG_araBAD_ilvBN (ilvBN-tunable E. coli)
0
E. coli K-12 BW25113 ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH (ilvIH-tunable E. coli)
0
E. coli K-12 BW25113 pSW3_lacI+ pACG_araBAD_ilvGM (ilvGM-tunable E. coli)
0
Main culture was prepared differently, depending on the tested cultivation conditions:
• Standard cultivation conditions
OD600nm at the end of the pre-cultivation was measured and a given volume was used to inoculate a 5
mL starting volume Pall Micro24 mini-reactor (Microreactor Technologies Inc.) so that initial OD600nm
was 0.4. The mini-reactor medium consisted of 1:3 supplemented TUB medium containing 4 g/L
glucose, 100 µg/mL ampicillin, 25 µg/mL chloramphenicol (only for tunable E. coli strains) and 1
µL/mL Desmophen antifoam. Medium was also supplemented with different concentrations of L-
arabinose. Cultivation was carried out at 37 °C and the pH was maintained at 7 by automatic control
with NH4OH and CO2. Stirrer speed was set to 800 rpm and DO set-point to 25 %, maintaining the last
by automatically increasing the oxygen flow into the mini-reactor. Batch phase lasted around 4h.
After batch phase was finished, 1 mL 400 g/L EnPump 200 solution and 50 µL 3000 U/L amylase
solution were manually added into the mini-reactor, hence starting the fed-batch phase. 30 min after
beginning of the fed batch phase, expression of recombinant mini-proinsulin was induced by manual
addition of an IPTG pulse to a final concentration of 0.5 mM. Fed-batch phase was active for 3.5 h.
• Cultivation conditions triggering ncBCAA formation
Cultivation was performed as described for the standard cultivation. However, immediately after
beginning of the fed-batch phase, a 0.833 g/L pyruvate pulse was manually added into the reactor.
During the following 5 min after pyruvate addition, DO set-point was set to 0, so that no oxygen was
supplied into the mini-reactor during that period, hence ensuring oxygen limitation. 30 min after the
first pyruvate pulse, expression of recombinant mini-proinsulin was induced by manual addition of an
IPTG pulse to a final concentration of 0.5 mM. After induction, sequential 0.833 g/L pyruvate pulses
were manually performed each 30 min as described above for a total of 5 pulses. Between pulses, DO
set-point was re-established to 25 %. Fed-batch phase was active for 3.5 h.
Cultivations were performed in 2 mini-reactor plates, comprising a total of 48 wells. Table 26 and
Table 27 show, for each tested cultivation, the well position, the tested strain, the L-arabinose
concentration employed as well as the cultivation conditions applied.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
115
Table 26. Overview of the different cultivation conditions tested in each well of the first mini-reactor plate with
strains E. coli K-12 BW25113 pSW3_lacI+, leuA-tunable E. coli, ilvC-tunable E. coli and thrA-tunable E. coli.
1
2
3
4
5
6
A
E. coli K-12
BW25113
pSW3_lacI+
0 % L-ara
Standard
E. coli K-12
BW25113
pSW3_lacI+
0 % L-ara
Standard
E. coli K-12
BW25113
pSW3_lacI+
0 % L-ara
Standard
E. coli K-12 BW25113
pSW3_lacI+
0 % L-ara
Pyruvate + O2 lim.
E. coli K-12 BW25113
pSW3_lacI+
0 % L-ara
Pyruvate + O2 lim.
E. coli K-12
BW25113
pSW3_lacI+
0 % L-ara
Pyruvate + O2 lim.
B
leuA-tunable E. coli
0.1 % L-ara
Standard
leuA-tunable E. coli
0.2 % L-ara
Standard
leuA-tunable E. coli
0.4 % L-ara
Standard
leuA-tunable E. coli
0.1 % L-ara
Pyruvate + O2 lim.
leuA-tunable E. coli
0.2 % L-ara
Pyruvate + O2 lim.
leuA-tunable E. coli
0.4 % L-ara
Pyruvate + O2 lim.
C
ilvC-tunable E. coli
0.4 % L-ara
Standard
ilvC-tunable E. coli
0.8 % L-ara
Standard
ilvC-tunable E. coli
1.6 % L-ara
Standard
ilvC-tunable E. coli
0.4 % L-ara
Pyruvate + O2 lim.
ilvC-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim.
ilvC-tunable E. coli
1.6 % L-ara
Pyruvate + O2 lim.
D
thrA-tunable E. coli
0.4 % L-ara
Standard
thrA-tunable E. coli
0.8 % L-ara
Standard
thrA-tunable E. coli
1.6 % L-ara
Standard
thrA-tunable E. coli
0.4 % L-ara
Pyruvate + O2 lim.
thrA-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim.
thrA-tunable E. coli
1.6 % L-ara
Pyruvate + O2 lim.
Table 27. Overview of the different cultivations tested in each well of the second mini-reactor plate with strains
E. coli BW25113 pSW3_lacI+, ilvIH-tunable E. coli, ilvA-tunable E. coli, ilvBN-tunable E. coli and ilvGM-tunable E.
coli.
1
2
3
4
5
6
A
E. coli K-12 BW25113
pSW3_lacI+
0 % L-ara
Standard
E. coli K-12 BW25113
pSW3_lacI+
0 % L-ara
Pyruvate + O2 lim.
ilvIH-tunable E. coli
0.05 % L-ara
Standard
ilvIH-tunable E. coli
0.8 % L-ara
Standard
ilvIH-tunable E. coli
0.05 % L-ara
Pyruvate + O2 lim
ilvIH-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim
B
ilvA-tunable E. coli
0.05 % L-ara
Standard
ilvA-tunable E. coli
0.2 % L-ara
Standard
ilvA-tunable E. coli
0.8 % L-ara
Standard
ilvA-tunable E. coli
0.05 % L-ara
Pyruvate + O2 lim
ilvA-tunable E. coli
0.2 % L-ara
Pyruvate + O2 lim
ilvA-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim
C
ilvBN-tunable E. coli
0.05 % L-ara
Standard
ilvBN-tunable E. coli
0.2 % L-ara
Standard
ilvBN-tunable E. coli
0.8 % L-ara
Standard
ilvBN-tunable E. coli
0.05 % L-ara
Pyruvate + O2 lim
ilvBN-tunable E. coli
0.2 % L-ara
Pyruvate + O2 lim
ilvBN-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim
D
ilvGM-tunable E. coli
0.05 % L-ara
Standard
ilvGM-tunable E. coli
0.2 % L-ara
Standard
ilvGM-tunable E. coli
0.8 % L-ara
Standard
ilvGM-tunable E. coli
0.05 % L-ara
Pyruvate + O2 lim
ilvGM-tunable E. coli
0.2 % L-ara
Pyruvate + O2 lim
ilvGM-tunable E. coli
0.8 % L-ara
Pyruvate + O2 lim
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
116
Results
5.8.2 Mini-proinsulin analysis by SDS-PAGE
Determination of recombinant mini-proinsulin concentration after induction for each strain under
each tested cultivation condition was carried out by densitometry analysis by SDS-PAGE exactly as
described in section 5.7.1.2 (protein gels not shown). Results are shown in Table S8 and Table S9,
corresponding to mini-reactor plates presented at Table 26 and Table 27, respectively.
5.8.3 ncBCAA analysis
Samples for amino acid analysis by GC-FID were prepared exactly as aforementioned in section
5.7.2.3. Concentrations of ncBCAA in both intracellular soluble protein fraction and inclusion body
fraction for each E. coli K-12 BW25113 ΔlgeneX pSW3_lacI+ pACG_araBAD_geneX mutant strain
under different L-arabinose concentrations and cultivation modes are shown in Figure 37 and Figure
38, respectively.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
117
Figure 37. Molar concentrations of norvaline (blue bars), norleucine (red bars) and β-methylnorleucine (green
bars) normalized to OD600nm in the intracellular soluble protein fraction from samples taken 3h after IPTG
induction of cultivations of ilvC-tunable E. coli (A), leuA-tunable E. coli (B), thrA-tunable E. coli (C), ilvIH-tunable
E. coli (D), ilvA-tunable E. coli (E), ilvBN-tunable E. coli (F) and ilvGM-tunable E. coli (G) in a 10mL mini-reactor
subjected to different L-arabinose concentrations and cultivation modes. Two cultivation modes were tested:
standard cultivation conditions (-) and conditions triggering ncBCAA production, i.e. pyruvate pulses combined
with O2 limitation (+). Strain E. coli K-12 BW25113 pSW3_lacI+ (indicated as “wild type E. coli" in the chart) was
employed as a control for comparison.
G
E F
C D
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
118
Results
Under standard cultivation conditions, for ilvC- and ilvIH-tunable E. coli strains, norvaline and
norleucine concentrations in the intracellular soluble protein fraction decreased when adding
increasing concentrations of L-arabinose into the medium. The opposite behavior was observed for
ilvBN-tunable E. coli. For thrA-, ilvA- and ilvGM-tunable E. coli strains, norvaline and norleucine
concentrations showed a reduction with respect to the control E. coli strain but L-arabinose
concentration did not seem to have a clear effect on ncBCAA concentration (Figure 37,-). The highest
reduction of norvaline concentration with respect to the control E. coli strain was reported when
inducing ilvIH- (-63.6%) and ilvGM-tunable E. coli strains (-75.9%) with 0.8 % L-arabionse. The highest
decrease of norleucine concentration was shown after inducing ilvIH- (-69.9%) and ilvGM-tunable E.
coli strains (-100%) with 0.8 % L-arabinose. The value of 100% did not correspond to a measured
norleucine concentration of 0. In that case norleucine concentration could simply not be measured
since it was lower than the detection limit of the GC.
Under standard cultivation conditions, for almost all tested strains and L-arabinose concentrations,
no significant variation was reported for β-methylnorleucine concentration in the intracellular
soluble protein fraction, with exception of ilvBN-tunable E. coli induced with 0.05 % L-ara, which
shown a reduction around 42%. In addition, effect of increasing L-arabinose concentrations did not
have a clear effect on β-methylnorleucine concentration (Figure 37,-).
Under cultivation conditions triggering ncBCAA formation (i.e. pyruvate pulses combined with O2
limitation), for ilvC- and leuA-tunable E. coli strains, norvaline and norleucine concentration in the
intracellular soluble protein fraction tended to decrease when adding increasing concentrations of L-
arabinose into the medium. The opposite behavior was observed for ilvBN- and ilvIH-tunable E. coli
strains. For thrA- and ilvA-tunable E. coli strains, norvaline and norleucine concentration showed a
reduction with respect to the control E. coli strain but L-arabinose concentration did not seem to
have a clear effect on ncBCAA concentration. For ilvGM-tunable E. coli, only norleucine concentration
reported a reduction with respect to the control E. coli strain. L-arabinose concentration did not
seem to have a clear effect on ncBCAA concentration in this case (Figure 37,+).
For the intracellular soluble protein fraction, similar results were obtained when expressing data as
ratio of the concentration of ncBCAA with respect to the canonical counterpart (Figure S74).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
119
Figure 38. Molar concentrations of norvaline (blue bars) and norleucine (red bars) normalized to OD600nm in the
inclusion body fraction from samples taken 3 h after IPTG induction of cultivations of ilvC-tunable E. coli (A),
leuA-tunable E. coli (B), thrA-tunable E. coli (C), ilvIH-tunable E. coli (D), ilvA-tunable E. coli (E), ilvBN-tunable E.
coli (F) and ilvGM-tunable E. coli (G) in a 10mL mini-reactor subjected to different L-arabinose concentrations
and cultivation modes. Two cultivation modes were tested: standard cultivation conditions (-) and conditions
A B
C D
E F
G
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
120
Results
triggering ncBCAA production, i.e. pyruvate pulses combined with O2 limitation (+). Strain E. coli K-12 BW25113
pSW3_lacI+ (indicated as “wild type E. coli” in the chart) was employed as a control for comparison.
Under standard cultivation conditions, ilvC- and ilvIH-tunable E. coli reported a significant decrease of
norvaline and norleucine concentrations in the inclusion body fraction when adding increasing
concentrations of L-arabinose into the medium. The opposite behavior was observed for ilvBN-
tunable E. coli. For thrA- and ilvGM-tunable E. coli strains, norvaline and norleucine concentrations
showed a significant reduction with respect to the control E. coli strain but effect of increasing L-
arabinose concentrations did not show a clear effect on ncBCAA concentration. For ilvA-tunable E.
coli no significant variation of norvaline and norleucine concentration was reported with respect to
the control E. coli strain and effect of increasing L-arabinose concentrations did not seem to show a
clear effect on ncBCAA concentration (Figure 38,-). The highest reduction of norvaline concentration
was reported when inducing ilvIH-tunable E. coli with 0.8 % L-arabionse (-40.7%) and ilvGM-tunable
E. coli with 0.05% L-arabinose (-61.8%). The highest decrease of norleucine concentration was shown
after inducing ilvIH- (-70.8%) and ilvGM-tunable E. coli (-100%) with 0.8 % L-arabinose. As mentioned
before, the value of 100% did not correspond to a measured norleucine concentration of 0. In that
case norleucine concentration could simply not be measured since it was lower than the detection
limit of the GC.
Under cultivation conditions triggering ncBCAA formation (i.e. pyruvate pulsing combined with
oxygen limitation), all tested strains with exception of leuA- and ilvIH-tunable E. coli showed the
same pattern described at the previous paragraph. Interesingly, the two aforementioned strains
reported the opposite effect than under standard cultivation conditions (Figure 38,+).
For the inclusion body fraction, similar results were obtained when expressing data as ratio of the
concentration of ncBCAA with respect to the canonical counterpart (Figure S75).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
121
5.9 Screening of potential ilvGM- and ilvIH-tunable E. coli strains in a 15 L
reactor under conditions triggering ncBCAA formation
According to section 5.8 strains E. coli K-12 BW25113 pSW3_lacI+ pACG_araBAD_ilvGM (ilvGM-
tunable E. coli) and E. coli K-12 BW25113 ΔlvIH pSW3_lacI+ pACG_araBAD_ilvIH (ilvIH-tunable E. coli)
induced with 0.8% L-arabinose showed the best performance among all screened mutants in a 10 mL
mini-reactor, since they reported the most significant reduction of ncBCAA mis-incorporation into
recombinant mini-proinsulin in comparison with the control non-engineered E. coli strain. The aim of
this experiment was to verify the performance of the aforementioned potential tunable E. coli strains
in a 15L reactor under cultivation conditions triggering formation of ncBCAA, i.e. pyruvate pulses and
oxygen limitation, in order to confirm its advantage as strain ensuring product quality. For
comparison, the control non-engineered E. coli host (E. coli K-12 BW25113 pSW3_lacI+) was also
cultivated.
5.9.1 Cultivation mode
5.9.1.1 Cultivation of E. coli K-12 BW25113 pSW3_lacI+ (control strain) under conditions
triggering ncBCAA formation
Cultivation operation was already described in section 5.7.2.1.2 (Figure 30).
5.9.1.2 Cultivation of ilvGM-tunable E. coli under conditions triggering ncBCAA formation
Cultivation operation was already described in section 5.7.2.1.2 (Figure 30) and only minor changes
were done in order to adapt the cultivation process to ilvGM-tunable E. coli strain. Both pre-culture
and reactor medium contained additionally 25 µg/mL chloramphenicol. The reactor medium
additionally contained 0.8% L-arabinose, necessary to induce expression of gene ilvGM hosted in
plasmid pACG_araBAD_ilvGM. The feeding solution was also additionally supplemented with 25
µg/mL chloramphenicol and 0.8% L-arabinose. A general overview of the cultivation is shown in
Figure 39.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
122
Results
C
A
B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
123
Figure 39. Overview of the cultivation of E. coli K-12 BW25113 pSW3_lacI+ pACG_araBAD_ilvGM (ilvGM-tunable
E. coli) in a 15L reactor under conditions triggering ncBCAA formation, during the whole cultivation process (A),
during fed-batch period (B) and during pyruvate pulsing (C). Different cultivation phases are shown in the
diagram as BE (first 2h of batch phase), KUL (12h cold period at 15 °C), FE (remaining 2h of batch phase), F1 (3h
exponential fed-batch phase) and IND (induction, linear fed-batch phase and pyruvate pulsing). IPTG induction
was performed for 30 minutes (20.5 to 21 h cultivation time). Pyruvate pulses are indicated by orange arrows
(C). Present in the diagram axes, Flow1 corresponds to the flow rate (g/h) of the pump transporting the feed
solution into the reactor while Flow2Y corresponds to one tenth of the flow rate (g/h) of the pump transporting
either the IPTG solution used for induction or the pyruvate solution for pulsing. Exponential fed-batch was, due
to a programming error, started shortly before glucose was completely consumed in the batch phase. Hence a
small glucose accumulation was reported at around 17h. Thus, exponential feeding was shortly shut down until
glucose was completely depleted and then, activated again.
5.9.1.3 Cultivation of ilvIH-tunable E. coli under conditions triggering ncBCAA formation
Cultivation operation was already described in section 5.7.2.1.2 (Figure 30) and only minor changes
were done in order to adapt the cultivation process to ilvIH-tunable E. coli. Both pre-culture and
reactor medium contained additionally 25 µg/mL chloramphenicol. The reactor medium additionally
contained 0.8% L-arabinose, necessary to induce expression of gene ilvIH hosted in plasmid
pACG_araBAD_ilvIH. The feeding solution was also additionally supplemented with 25 µg/mL
chloramphenicol and 0.8% L-arabinose. A general overview of the cultivation is shown in Figure 40.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
124
Results
A
B
C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
125
Figure 40. Overview of the cultivation of E. coli K-12 BW25113 ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH (ilvIH-
tunable E. coli) in a 15L reactor under conditions triggering ncBCAA formation, during the whole cultivation
process (A), during fed-batch period (B) and during pyruvate pulsing (C). Different cultivation phases are shown
in the diagram as BE (first 2h of batch phase), KUL (12h cold period at 15 °C), FE (remaining 2h of batch phase),
F1 (3h exponential fed-batch phase) and IND (induction, linear fed-batch phase and pyruvate pulsing). IPTG
induction was performed for 30 minutes (19.9 to 20.4 h cultivation time). Pyruvate pulses are indicated by
orange arrows (C). Present in the diagram axes, Flow1 corresponds to the flow rate (g/h) of the pump
transporting the feed solution into the reactor while Flow2Y corresponds to one tenth of the flow rate (g/h) of
the pump transporting either the IPTG solution used for induction or the pyruvate solution for pulsing.
Exponential fed-batch was, due to a programming error, started shortly before glucose was completely
consumed in the batch phase. Hence a small glucose accumulation was reported at around 16.5hh. Thus,
exponential feeding was shortly shut down until glucose was completely depleted and then, activated again. At
aroung 23.5h glucose suddenly started being accumulated. It was hypothesized that the pump was going too
fast even though Flow1 values were correct. Hence, pump was shortly shut down and restarted. However, after
reaching the pump again the set Flow1 value, glucose still kept accumulating. This might be explained due to a
problem inherent to the cultivated strain, i.e. culture is reaching dead phase.
5.9.2 Mini-proinsulin analysis by HPLC
Recombinant mini-proinsulin concentration from hourly samples was analyzed according to an in
house HPLC method available at Sanofi-Aventis Deutschland GmbH. 1 mL of culture broth was
directly used for analysis. PPI concentrations resulting from the in house HPLC analytical method are
shown in Figure 41. PPI concentrations determined for all three tested E. coli strains were really
similar. In addition, cell growth behavior (represented by OD600nm and CDW) was also comparable for
all tested strains (Figure S76), suggesting that specific production of recombinant mini-proinsulin
remains the same, independently of the strain employed.
Figure 41. Estimated recombinant mini-proinsulin concentrations over time after induction of different E. coli
cultivations in a 15 L reactor under conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and
oxygen limitation (PYR-O2). Indicated in the legend, “WT E.coli” refers to the wild type strain E. coli K-12
BW25113 pSW3_lacI+, “ilvGM-tunable E. coli” alludes to strain E. coli K-12 BW25113 pSW3_lacI+
pACG_araBAD_ilvGM and “ilvIH-tunable E. coli” corresponds with strain E. coli K-12 BW25113 ΔilvIH
pSW3_lacI+ pACG_araBAD_ilvIH.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
126
Results
5.9.3 Analysis of ncBCAA
Samples for amino acid analysis by GC-FID were prepared exactly as aforementioned in section
5.7.2.3. Concentrations of ncBCAA present in the intracellular soluble protein fraction and in the
inclusion body fraction over cultivation time for each tested strain are shown in Figure 42 and Figure
43, respectively.
Figure 42. Molar concentrations of norvaline (A, B), norleucine (C, D) and β-methylnorleucine (E, F) normalized
to OD600nm present in the intracellular soluble protein fraction calculated over time after induction of different
E. coli cultivations in a 15L reactor under cultivation conditions triggering ncBCAA accumulation, i.e. pyruvate
pulsing and oxygen limitation (PYR-O2), in a big (A, C, E) and reduced y-axis scale (B, D, F). Indicated in the
legend, “WT E.coli” refers to the wild type strain E. coli K-12 BW25113 pSW3_lacI+, “ilvGM-tunable E. coli”
alludes to strain E. coli K-12 BW25113 pSW3_lacI+ pACG_araBAD_ilvGM and “ilvIH-tunable E. coli” corresponds
with strain E. coli K-12 BW25113 ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH. Orange arrows indicate time points
where 1 g/L pyrvate pulse combined with 5 min O2 limitation was applied.
The cultivation of the control E. coli strain subjected to pyruvate pulses combined with O2 limitation
(“WT E. coli, PYR-O2”) reported a progressive accumulation of norleucine and β-methylnorleucine in
A B
C D
E F
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
127
the intracellular soluble protein fraction over time after induction, being that more significant for
norleucine. Furthermore, norvaline concentration also increased progressively under
aforementioned cultivation conditions, but only until 3h after induction. From that time point on,
norvaline concentration progressively dropped until reaching initial values at 5 h after induction. This
might suggest that, after 2h from last pyruvate pulse combined with O2 limitation, its associated
effect triggering norvaline accumulation is not active anymore (Figure 42).
Both tested potential mutants in cultivations “ilvGM-tunable E. coli, PYR-O2” and “ilvIH-tunable E.
coli, PYR-O2” reported a dramatic reduction of norvaline and norleucine concentrations in the
intracellular soluble protein fraction, being such decrease higher for norleucine in “ilvGM-tunable E.
coli, PYR-O2”. However, β-methylnorleucine concentrations did not significantly vary with respect to
the control cultivation. It is noteworthy to highglight that, for most samples, norvaline could not be
properly detected since concentrations were under the limit of detection of the GC-FID equipment
(Figure 42).
Figure 43. Molar concentrations of norvaline (A, C) and norleucine (B, D) normalized to OD600nm present in the
inclusion body fraction calculated over time after induction of different E. coli cultivations in a 15L reactor
under cultivation conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-
O2), in a big (A, B) and reduced y-axis scale (C, D). Indicated in the legend, “WT E.coli” refers to the wild type
strain E. coli K-12 BW25113 pSW3_lacI+, “ilvGM-tunable E. coli” alludes to strain E. coli K-12 BW25113
pSW3_lacI+ pACG_araBAD_ilvGM and “ilvIH-tunable E. coli” corresponds with strain E. coli K-12 BW25113
ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH. Orange arrows indicate time points where 1 g/L pyruvate pulse
combined with 5 min O2 limitation was applied.
The cultivation of the control E. coli strain subjected to pyruvate pulses combined with O2 limitation
(“WT E. coli, PYR-O2”) reported a progressive accumulation of norvaline and norleucine in the
inclusion body fraction over time after induction. Again, and similar to reported in the intracellular
soluble fraction, both tested potential mutants in cultivations “ilvGM-tunable E. coli, PYR-O2” and
“ilvIH-tunable E. coli, PYR-O2” reported a dramatic reduction of norvaline and norleucine
A B
C D
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
128
Results
concentrations in the inclusion body fraction, being this decrease even higher for norleucine in
“ilvGM-tunable E. coli, PYR-O2”. Norvaline could not be detected in any case for both tested mutants.
β-methylnorleucine could not be detected in any tested samples (Figure 43).
Similar results are also shown for the inclusion body fraction when normalizing amino acid
concentration to PPI mass (Figure S77).
5.9.4 Acetate and formate analysis
Acetate and formate concentrations from hourly samples were offline analysed by in house
enzymatic assays available at Sanofi-Aventis Deutschland GmbH. Sample preparation for acetate and
formate analysis was as follows: 10 mL of culture broth were inactivated with Bardac, centrifuged at
4000 g for 10 min and resultant supernatant was decanted and filtered through a 0.45 µm filter. The
filtrate was then used for analysis. Estimated concentrations of acetate and formate in culture broth
supernatants over cultivation time are shown in Figure 44.
B
A
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Results
129
Figure 44. Concentration of acetate (A) and formate (B) present in culture broth supernatants calculated over
fed-batch time of different E. coli cultivations in a 15L reactor under conditions triggering ncBCAA
accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-O2). Indicated in the legend, “WT E. coli” refers
to the wild type strain E. coli K-12 BW25113 pSW3_lacI+, “ilvGM-tunable E. coli” alludes to strain E. coli K-12
BW25113 pSW3_ lacI+ pACG_araBAD_ilvGM and “ilvIH-tunable E. coli” corresponds with strain E. coli K-12
BW25113 ΔilvIH pSW3_ lacI+ pACG_araBAD_ilvIH. Orange arrows indicate time points where 1 g/L pyruvate
pulse combined with 5 min O2 limitation was applied. Blue arrow points out time point where IPTG induction
was carried out.
For all tested cultivations, a similar acetate profile was observed over time. Acetate concentration
reached a maximum after batch phase was finished (0 h fed-batch time). Afterwards, acetate levels
decreased to a minimum around 0.2 g/L at about 4h after induction. From that time on, acetate
concentration rapidly started to increase as a consequence of the metabolic alteration triggered by
pyruvate pulsing combined with O2 limitation. Once pulsing was finished, acetate levels started to
slowly decrease until the end of the cultivation, except for cultivation “ilvIH-tunable E. coli, PYR-O2”,
where acetate still increased (Figure 44). This is due to the problems of glucose accumulation
reported at the end of this cultivation (Figure 40).
For all tested cultivations, a similar formate profile was observed over time as well. Formate levels
remained close to 0 until about 4h after induction. However, from that time point on, formate
concentration progressively increased as a consequence of the metabolic alteration triggered by
pyruvate pulsing combined with O2 limitation. Once pulsing was finished, formate levels stopped
increasing, reaching a plateau (Figure 44).
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
130
Discussion
6. Discussion
6.1 Analysis of mini-proinsulin expression in E. coli K-12 BW25113 containing
different variants of plasmid pSW3
Expression of mini-proinsulin from plasmid pSW3 is under the control of a Ptac promoter. Ptac is a
hybrid promoter where -35 region is derived from the trp promoter while -10 region comes from
the lac UV5 promoter (de Boer et al., 1983). The Ptac promoter is inducible by IPTG addition but,
under certain conditions, it is also leaky, i.e. expression is also triggered in a basal level even when no
inducer is present. This is because the LacI repressor does not tightly bind to the operator and can
dissociate. Hence, when using lac-based promoters in cloning vectors is important to take into
consideration that enough LacI repressor is present in order to avoid leaky expression of the
promoter. The number of LacI molecules necessary to repress a lac-based promoter mainly depends
on the number of operators regulating the promoter, the affinity of the repressor to the operator
and the copy number of the plasmid bearing the lac promoter (Penumetcha et al., 2010). Different
variants of lacI promoter sequences leading to different expression levels of lacI have been reported
in E. coli strains: lacI+ and lacIq in Calos (1978), I-UJ177 in Calos and Miller (1980) and lacIq1 in Calos
and Miller (1981). According to Glascock et al. (1998), an E. coli wild type cell (lacI+ variant) has about
10 LacI molecules. The lacIq promoter version shows 10-fold expression enhancement of lacI if
compared with the lacI+ variant (100 LacI molecules) while expression by lacIq1 is 17-fold stronger
than that of lacIq and 170-fold stronger than that of lacI+ (1700 LacI molecules).
Plasmid pSW3 does not contain lacI and it was especially designed for expression in strain E. coli K-12
W3110M. This strain is lacIq and it expresses endogenously 10-fold more LacI repressor than a lacI+
strain, hence allowing a tight expression of plasmid pSW3. However, in this study, E. coli K-12
BW25113 was used as a model strain. Results reported in section 5.1 revealed that, as opposed to E.
coli K-12 W3110M, Ptac promoter controlling expression of mini-proinsulin in plasmid pSW3 is leaky in
the genetic background of E. coli K-12 BW25113, being no significant difference in mini-proinsulin
expression level between induced and non-induced samples. This suggests that E. coli K-12 BW25113
is a lacI+ strain so that there is not sufficient endogenous LacI repressor to maintain Ptac promoter
inactive when no IPTG induction is performed. However, plenty of literature wrongly stated that
E.coli K-12 BW25113 was lacIq (reviewed by Baba et al., 2006, Supplementary table 1). Moreover, the
sequence of the 3 allelic variants of the lacI promoter (lacI+, lacIq and lacIq1) were aligned to the lacI
promoter located in the genome of E. coli K-12 BW25113 and it was confirmed that strain is actually
lacI+ (Figure S78).
Observed results in section 5.1 are in accordance with the investigation of Glascock and Weickert
(1998), where a similar scenario was described. They analyzed the recombinant protein expression
behavior of the lacI+ E. coli strain SGE1661 containing the medium-copy plasmid pSGE714 (no lacI),
which expresses recombinant β-galactosidase under the control of a Ptac promoter. They showed that
non-induced cells already express around half of the total recombinant β-galactosidase produced by
cells induced with 0.1 mM IPTG. However, it is noteworthy to point out that: (i) unlike plasmid
pSGE714, plasmid pSW3 is a high-copy plasmid, since it contains the pBR22 origin of replication
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Discussion
131
without the rop gene; (ii) instead of the 0.1 mM IPTG used in Glascock and Weickert (1998), 0.5 mM
IPTG was used in the current study. These two considerations would explain why no significant
differences in mini-proinsulin expression levels were reported between induced and non-induced
samples.
An E. coli K-12 BW25113 cell, as lacI+, would have approximately 10 LacI molecules, which is not
sufficient to repress Ptac promoter present in pSW3 because there are 115 plasmid copies per cell (or
115 pSW3 operator sites per cell, as there is a single operator in the Ptac promoter) (Glascock and
Weickert, 1998). An E. coli K-12 W3110M cell, as lacIq, would have approximately 100 LacI molecules
and it should be sufficient to repress the Ptac promoter. Considering that strain E. coli K-12 BW25113
(lacI+) expressing a pSW3 variant contains about 115 copies of the plasmid, expression levels of lacI
are predicted to be 0, 1150 and 11500 LacI molecules for pSW3, pSW3_lacI+ and pSW3_lacIq,
respectively (Table 28). Those estimations are in accordance with results reported in section 5.2. As
opposed to plasmid pSW3, plasmid variants pSW3_lacI+ and pSW3_lacIq did not report promoter
leakiness, confirming sufficient LacI levels. However, weak expression was shown for plasmid
pSW3_lacIq, which could be explained due to the high content of LacI repressor expressed. Hence, it
may be that the concentration of IPTG used for induction in that specific case was not high enough to
effectively trigger expression of the recombinant protein. It might also be that the high amount of
LacI being expressed reduced the expression efficiency of the recombinant protein due to an
overexploitation of the transcription and translation cell machinery.
All in all, research hypothesis 7 could be confirmed.
Table 28. Estimated number of LacI molecules expressed by two different E. coli K-12 strains expressing
different pSW3 plasmid variants. Data obtained from Glascock and Weickert (1998).
Strain
Estimated number of LacI
molecules expressed in the
plasmid (115 copies/cell)
Estimated number of LacI
molecules expressed in the
genome
E. coli K-12 W3110M (lacIq)
0
100
E. coli K-12 W3110M (lacIq) pSW3
0
100
E. coli K-12 BW25113 (lacI+)
0
10
E. coli K-12 BW25113 (lacI+) pSW3
0
10
E. coli K-12 BW25113 (lacI+) pSW3_lacI+
1150
10
E. coli K-12 BW25113 (lacI+) pSW3_lacIq
11500
10
6.2 Evaluation of L-arabinose induction in E. coli BW25113 ΔgeneX
expressing pSW3_lacI+ and pACG_araBAD_geneX (geneX-tunable E. coli)
After screening of various L-arabinose concentrations, the effective induction range of L-arabinose
for plasmid pACG_araBAD_geneX was determined to be 0.05-1.6 % (section 5.3). This concentration
range is higher than the one normally used for commercial plasmids based in the araBAD promoter.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
132
Discussion
For instance, induction of the commercial plasmid pBAD_DEST49 is recommended within the range
0.00002-0.2 % L-arabinose (Website 4). Variation in the L-arabinose amounts necessary for induction
might be explained due to differences in the selected origin of replication, which determines the
plasmid copy number in the cell. While plasmid pBAD_DEST49 contains a pUC origin, acting as a high
copy plasmid, plasmid pACG_araBAD_geneX contains an F plasmid-based origin of replication,
ensuring 1-copy plasmid. The selection of araBAD promoter variants with enhanced transcriptional
strength (Website 5) would have probably resulted in a reduction of the effective L-arabinose
induction range.
According to results reported in section 5.3, it is noteworthy that, for each target gene, different
induction strength, i.e. L-arabinose concentration, was necessary in order to trigger genetic
expression levels enough to recover cell growth levels of the control E. coli strain. This might be
explained due to the different degree of essentiality of the target genes for the cell. The L-arabinose
concentration needed to fully recover growth levels of the control E. coli strain was especially high
for the thrA-tunable E. coli strain (0.4 %), followed by ilvC-tunable E. coli (0.4 %), leuA-tunable E. coli
(0.1 %) and ilvA-tunable E. coli (0.05 %). Moreover, when no L-arabinose was supplemented, cell
growth reported for the aforementioned E. coli mutants was close to 0. These results are in
accordance with the study of Baba et al. (2006). They generated numerous single gene knock-out E.
coli mutants and then tested growth behavior (OD600nm) of the different mutants 24 and 48 h after
cultivation in LB and minimal medium (MOPS). In that study, the aforementioned genes were
classified as non-essential since mutants were able to grow in LB medium. However, growth in
minimal medium was reported to be almost 0.
For ilvIH-, ilvBN- and ilvGM-tunable E. coli strains, no significant differences in growth behavior
between non-induced and induced samples were reported, independently on the L-arabinose
concentration employed for induction (section 5.3). As a consequence, differently from reported for
mutants mentioned in the previous paragraph, induction efficiency of L-arabinose could not be
evaluated by monitoring cell growth. These results suggest that lack of one of the three ilvGM, ilvIH
and ilvBN gene products does not dramatically affect growth behavior since, as isoenzymes, its
function can be also carried out by the other two available gene products. These results are also in
accordance with the study of Baba et al., (2006). They showed that ilvIH-, ilvBN- and ilvGM mutants
were also able to grow in minimal medium.
6.3 Evaluation of m-toluate induction in E. coli BW25113 ΔgeneX expressing
pSW3_lacI+ and pACG_XylSPm_geneX (geneX-tunable E. coli)
As demonstrated in section 5.4, addition of m-toluic acid does not have an effect on growth behavior
if compared with the non-induced case for any tested strains. This might be explained due to non-
sufficient induction strength of the selected XylSPm/Pm promoter variant. The unmodified wild type
Pm promoter variant was selected for this study since it was demonstrated to have a less basal
expression than other mutagenized high-level expression variants (Balzer et al., 2013; Binder et al.,
2016). However, this promoter variant was successfully tested mainly in medium copy plasmids (40-
60 copies per cell) containing the RK2 origin of replication (i. e. plasmid pJB658, Blatny et al., 1997).
In the current study, however, the selected plasmid for tunable expression contains an F plasmid-
based origin of replication, ensuring 1-copy plasmid, and this might explain the non-sufficient
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Discussion
133
induction of the Pm promoter. The generation of new pACG_XylSPm_geneX plasmid variants
containing other mutagenized high-level expression variants of the Pm promoter (Winther-Larsen et
al., 2000) might improve the induction strength. Nevertheless, since induction of the tunable plasmid
pACG_araBAD_geneX was demonstrated to properly work, this variant was selected for gene
regulation and no further investigation was performed with the pACG_XylSPm_geneX plasmid
variant. In addition, it can be concluded that 10 mM m-toluic acid triggers cellular toxicity since under
this concentration OD600nm reached almost 0 for all tested strains. This is not in accordance with many
literature reports stating as an advantage of the XylSPm/Pm promoter system the non-toxic
character of m-toluic acid for E. coli cells (Binder et al., 2016).
6.4 Establishment of a GC-FID method allowing analysis of canonical and
non-canonical amino acids
Among the analyzed amino acids, β-methylnorleucine was the only one represented by two different
peaks in the chromatogram (Table 24, section 5.5). Since β-methylnorleucine contains two
asymmetric carbons, 4 different configurations, i.e. 4 stereoisomers, are possible (SS, SR, RS and RR).
Configurations SS and RR and configurations SR and RS are enantiomers and have the same chemical
and physical properties. Hence, when injecting a β-methylnorleucine solution containing the 4
stereoisomers into the GC for analysis, two peaks are generated due to variations in the retention
time of each pair of enantiomers. The β-methylnorleucine standard employed for calibration
contained the four variants. However, according to Muramatsu et al. (2002), the final stereochemical
configuration of the β-methylnorleucine produced by E. coli is predicted to be 2S,3S because this is
also the configuration of the natural L-isoleucine biosynthesized by E. coli. Hence, only the first peak
at 3.398 min was considered for β-methylnorleucine calibration and analysis.
Concerning the evaluation of acid hydrolysis effect on amino acid analysis (Figure 18, section 5.5), a
signal reduction was reported after hydrolysis for most of the analyzed amino acids, being this
especially evident for norvaline, norleucine, threonine, methionine and tyrosine. In addition, amino
acids asparagine, glutamine and tryptophan could not be measured after hydrolysis and the signal of
glutamate and aspartate reported a significant increase. These results are in accordance with
previous investigations such as the one from Pickering and Newton (1990) and the one from
Davidson (1997). According to Davidson (1997), serine and threonine are recovered in low yield due
to ester formation and due to the modification by dehydratin of the hydroxyl group present in the
side chain. Tyrosine is also generated in lower yields after hydrolysis due to modification of the
phenolic group in the side chain. Methionine is as well obtained in low yields after hydrolysis due to
oxidation of the thioether group present in the side chain. Tryptophan is not quantifiable since the
the indole group of the molecule is destroyed. Asparagin and glutamin are deamidated to the
respective acids (aspartate and glutamate). In addition, recovery of some hydrophobic amino acids
such as valine, isoleucine, leucine and alanine is also poor due to the high stability of the bonds
generated between those mentioned amino acids.
Depending on the target amino acid, different strategies have been described in order to diminish
effects of acid hydrolysis in the amino acid recovery yield: addition of reducing agents (for
methionine and tryptophane), drying of hydrolysates (for serine and threonine), addition of phenol
(for tyrosine) and/or extension of hydrolysis time (for valine, isoleucine, leucine and alanine), among
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
134
Discussion
others (Davidson, 1997). In this study, it was demonstrated that the extension of acid hydrolysis from
24h to 72h slighty improved recovery yield of most of the analyzed amino acids, reaching values close
to the theoretical ones (Figure 19, section 5.5). However, due to practical reasons and the fact that
improvement of hydrolysis performance was not dramatic, 24h was finally selected as the standard
time for acid hydrolysis in this study.
6.5 Establishment of cultivation conditions based on pyruvate pulsing and O2
limitation leading to an increase of ncBCAA mis-incorporation into
recombinant mini-proinsulin in E. coli
During fermentation in large-scale reactors, gradient zones of substrate, dissolved oxygen, pH and
other parameters are formed due to inefficient mixing and E. coli cells respond to these
environmental changes by modulating their metabolism (Schweder et al., 1999). For instance, E. coli
responds to glucose excess and oxygen limitation by shifting metabolism from oxidative respiration
to mixed-acid fermentation, resulting in overflow metabolism (Enfors et al., 2001). Under these
conditions, not only the mixed-acid fermentation products accumulate, but also pyruvate (Soini et
al., 2008, I). Pyruvate excess present intracellularly increases the metabolic flux going to ncBCAA
biosynthesis through the sequential keto acid chain elongation from pyruvate to α-ketocaproate over
α-ketobutyrate and α-ketovalerate by the actuation of the leu operon-encoded enzymes (Apostol et
al., 1997). This hypothesis is supported by the observations reported by Soini et al. (2011): the
combination of oxygen limitation with a constant glucose supply in a two-compartment STR-PFR
scale-down reactor showed a significant impact on enhancing norvaline biosynthesis due to pyruvate
accumulation in a recombinant E. coli cultivation. Furthermore, Soini et al. (2008, I) originally
reported accumulation of pyruvate-based amino acids such the ncBCAAs norleucine and norvaline as
well as alanine and valine in a standard STR fed-batch E. coli cultivation under glucose excess and
induced oxygen limitation upon a stirrer downshift. As mentioned before, concentration gradients
happening in large industrial-scale reactors due to deficient mixing can be also simulated in scale-
down reactors in the laboratory. In this investigation large-scale effects are reproduced in different
fermentation systems by combining pyruvate pulsing and O2 limitation. This novel cultivation
strategy might represent more accurately the physiological behavior of bacterial cultivations taking
place in large scale reactors.
Research hypothesis 2 was confirmed according to results reported in Figure 20 and Figure 28 (and
additionally, Figure S71 and Figure S72). The cultivation strategy consisting of combining pyruvate
pulsing and O2 limitation was demonstrated to be an alternative method to represent large-scale
effects in both shake flask and mini-reactor scales since concentrations of norleucine and norvaline
increased by using the aforementioned strategy. Similar results concerning increase of ncBCAA
biosynthesis under those cultivation conditions were also observed in a 15L reactor scale (Figure 32
and Figure 33). In this case, overflow metabolism triggered by pyruvate pulsing and O2 limitation
could also be demonstrated by analyzing concentrations of acetate and formate (Figure 34). As
shown in Figure 20, it could also be demonstrated that the triggered ncBCAA biosynthesis is directly
proportional to the amount of supplemented pyruvate.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Discussion
135
According to the defined research hypothesis 3, the use of O2 limitation in combination with
pyruvate pulses to mimic large-scale effects may have an advantage with respect to other strategies
based on glucose pulses. Pyruvate is the core substrate of the metabolic pathway leading to ncBCAA
biosynthesis. Hence, when addying pyruvate to the cultivation medium, this enters the cell and
accumulates intracellularly, rapidly triggering ncBCAA formation through the sequential keto acid
chain elongation by the actuation of the enzymes encoded by the leu operon. Unlike pyruvate, when
using glucose pulses, large-scale effects (i. e. increase of ncBCAA biosynthesis) can first be reported
once glucose is converted to pyruvate and this starts accumulating in the cell. In this study, the
intracellular soluble fraction of an E. coli cultivation reported the first significant increase of
norleucine and norvaline concentrations around 1.5h after triggering pyruvate pulses and oxygen
limitation (Figure 32). In a similar study, Soini et al. (2008, I) induced oxygen limitation in an E. coli
cultivation at a high glucose concentration. In that study, the first significant increase of norvaline
levels was reported around 1h after applying the aforementioned cultivation conditions, which is
even earlier than reported in this study by using pyruvate pulses (1.5h). This is not in agreement with
what was previously argued, being research hypothesis 3 rejected. This might be due to the fact that
the employed glucose concentration in the study of Soini et al. (2008, I) is much higher than the
pyruvate concentration used in this study, which was a total of 5 g/L. That might also be explained
due to the the presence in E. coli of different uptake/transport systems for both carbon substrates
and its regulation. Kreth et al. (2013) suggested the existence of at least three transport systems for
pyruvate in order to equilibrate intracellular pyruvate concentrations: an inducible uptake system
(Usp system), a constitutive uptake system (PrvT system) and an excretion system. The first two
systems were demonstrated to be controlled by catabolite repression. Hence, it might be that
glucose remains present in the cultivation medium during pyruvate pulsing inhibit transport of
pyruvate into the cell or that pyruvate accumulation into the cell is reverted by activation of the
pyruvate excretion system. Glucose can also be uptaken by means of different transport systems: the
glucose PTS (phosphotransferase system) system, the mannose PTS system, the galactose ABC
transporter, the galactose permease and the maltose ABC transporter (Steinsiek and Bettenbrock,
2012).
Furthermore, the inclusion body fraction of E. coli cultivation showed the first significant increase of
ncBCAA levels around 3.5h after triggering pyruvate pulses and oxygen limitation (Figure 33), which is
2h later than reported in the intracellular soluble protein fraction. This suggests that metabolic
effects triggered by pyruvate pulsing combined with O2 limitation occur first in the cytosol
(intracellular soluble protein fraction), where ncBCAAs are synthesized. Afterwards, translation
machinery would mis-incorporate those previously synthesized ncBCAA present in the cytosol into
the nascent recombinant proteins (inclusion body fraction). This is also supported by the fact that
both norvaline and norleucine present a higher concentration in the intracellular soluble fraction
than in the inclusion bodies. Accordingly, research hypothesis 4 was confirmed.
Unlike norvaline and norleucine, concentration of β-methylnorleucine did not significantly change
after exposing an E. coli cultivation to pyruvate pulsing and O2 limitation with respect to the non-
altered cultivation at both tested scales (Figure 32). This suggests that the metabolic effect triggered
by pyruvate pulsing and O2 limitation has a limited influence on β-methylnorleucine synthesis, in
contrast to norleucine and norvaline. The fact that β-methylnorleucine is the ncBCAA, the
biosynthetic pathway of which is the longest starting from pyruvate, also supports this hypothesis.
Norvaline and norleucine biosynthesis require the enzymes of the leu operon while β-
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
136
Discussion
methylnorleucine biosynthesis demands the additional actuation of the enzymes of the ilv operon.
Observed results might as well suggest that β-methylnorleucine is synthesized by an independent
alternative metabolic pathway.
In order to proof research hypothesis 5, the relative proportion of the three target ncBCAA was
evaluated taking into consideration the different analysed protein fractions, cultivation conditions
and tested scale. At shake flask level and mini-reactor scale, norvaline and β-methylnorleucine
concentrations reported in the intracellular soluble fraction under both standard cultivation
conditions and under pyruvate pulsing and O2 limitation conditions were higher than norleucine.
However, at 15 L reactor scale, norvaline and norleucine reported higher concentrations with respect
to β-methylnorleucine in the intracellular soluble fraction under both tested cultivation conditions.
As opposed to intracellular soluble fraction, analysis of ncBCAA in the inclusion body fraction at both
mini-reactor and 15L reactor scale under both tested cultivation conditions revealed that norleucine
concentration is significantly higher than norvaline and that β-methylnorleucine could not be
detected. This result is interesting taking into account that the expressed recombinant protein
contains only 3 methionines while 14 leucines and that norleucine and norvaline are analog to
methionine and leucine, respectively. Hence, a higher mis-incorporation of norvaline would be
actually expected. These observations suggest that the amount of a certain cBCAA in the amino acid
sequence of a recombinant protein is not the main factor determining mis-incorporation of the
respective ncBCAA analog in such recombinant protein. Accordingly, research hypothesis 5 was
rejected. In addition, these results might lead to the assumption that, despite being norvaline and
norleucine present in similar concentration levels intracellularly, the probability of
misaminoacylation by met-tRNA is higher than by leu-tRNA. Furthermore, in both inclusion body and
intracellular soluble fraction, norvaline and norleucine concentrations were higher at 15 L reactor
than at mini-reactor scale. Although not comparable to industrial scale reactors, it might be that
mixing in the 15 L reactor is inefficient, thus triggering formation of O2 and glucose gradients, which,
in turn, lead to ncBCAA formation, even under standard cultivation conditions.
The effect of recombinant protein induction in ncBCAA biosynthesis was also investigated. According
to reported results in Figure 33, IPTG-mediated induction of recombinant protein expression is the
main factor triggering ncBCAA formation under standard cultivation conditions, hence confirming
research hypothesis 6. The recombinant mini-proinsulin expressed in this study consisted of a total of
96 amino acids, 14 of which were leucine residues (14.5%) while an average E. coli protein only
contains 8.4% leucine (Neidhardt and Umbarger, 1996). Overexpression of leucine-rich recombinant
proteins cause depletion of the intracellular leucine pool which, in turn, causes de-regulation of the
enzymes encoded by the leu operon in order to counteract leucine limitation (Bogosian et al., 1989;
Apostol et al., 1997). In parallel, activation of the leu operon triggers ncBCAA biosynthesis through a
number of sequential chain elongation reactions starting from pyruvate (Sycheva et al., 2007).
Furthermore, under amino acid starvation conditions such as the aforementioned leucine depletion,
transcriptional regulation of numerous genes governing the BCAA biosynthetic pathway is driven by
the global RelA/SpoT modulon. Hence, operons leuABCD, ilvGMEDA and ilvBN as well as genes ilvC
and thrA are up-regulated whereas operon ilvIH is down-regulated by the global (p)ppGpp regulator
after amino acid starvation (Traxler et al., 2008; Umbarger, 1996; Tedin and Norel, 2001; Baccigalupi
et al., 1995). Considering this regulatory configuration and taking into account that ilvG is not active
in the E. coli strain used in this study (E. coli K-12 BW25113), metabolic flux would be redirected to
leucine, valine and ncBCAA biosynthesis.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Discussion
137
6.6 Screening of geneX-tunable E. coli strains
Research hypothesis 1 was defined by taking into consideration the BCAA metabolic pathway. The
following strategies were pursued by modulating regulation of target genes in E. coli in order to
reduce ncBCAA biosynthesis: (i) limit conversion of pyruvate to α-ketobutyrate, (ii) limit
transformation of threonine to α-ketobutyrate and (iii) limit conversion of α-ketobutyrate to α-
ketovalerate. Strategy (i) might be achieved by down-regulating operon leuABCD but also by up-
regulating operon ilvBN, strategy (ii) may be realised by down-regulating the thr genes as well as ilvA.
Strategy (iii) could be accomplished by down-regulating operon leuABCD and up-regulating ilvIH,
ilvGM and ilvC. According to this hypothesis novel E. coli strain mutants were genetically engineered
so that the expression of single target genes (leuA, thrA, ilvA, ilvC, ilvIH, ilvBN and ilvGM) could be
modulated in order to evaluate the effect of genetic modulation in ncBCAA biosynthesis. Screening of
the engineered E. coli mutants was performed in a 10 mL mini-reactor under standard cultivation
conditions. Additional screening of mutants under cultivation conditions triggering ncBCAA, i. e.
pyruvate pulses and O2 limitation, allowed elucidating performance of mentioned E. coli mutants
under large-scale cultivation conditions.
In order to support discussion of the reported results following hypothesis already described in
literature were assumed: (i) since the recombinant protein expressed in this study has a high leucine-
content, a depletion of the intracellular leucine pool after IPTG induction is expected, thus causing
de-regulation of the leu operon, which, in turn, triggers a relative increase of ncBCAA biosynthesis
(Bogosian et al., 1989; Apostol et al., 1997); (ii) under cultivation conditions based on pyruvate pulses
and O2 limitation an intracellular accumulation of pyruvate and, consequently, an increase of the
metabolic flux through leucine, valine and ncBCAA pathways is expected (Soini et al., 2008, I). In
addition, for each scenario, transcriptional and post-translational regulation of target genes
(summarized in Table 1, section 2.2.5) was also taken into consideration for discussion.
Down-regulation of leuA was hypothesized to limit ncBCAA biosynthesis by restricting the succesive
keto acid chain elongation reactions starting from pyruvate until ncBCAA formation. Some
investigations previously demonstrated that knocking-out one ore more leu genes reduces ncBCAA
biosynthesis (Fenton et al., 1994; Bogosian et al., 1989). Results obtained for leuA-tunable E. coli
under standard cultivation conditions in this study are in accordance with the logic of the metabolic
pathway since, for the mentioned E. coli mutant, norvaline and norleucine concentrations present in
both tested protein fractions progressively increased by adding increasing concentrations of L-
arabinose into the medium, i.e. by increasing leuA expression. Interestingly, under cultivation
conditions subjected to pyruvate pulses and O2 limitation, increasing leuA expression did not
translate into significant variation of ncBCAA concentrations but those remained always higher than
under standard cultivation conditions. This observation might be explained due to a saturation of the
metabolic pathway comprising the consecutive keto acid chain elongation reactions starting from
pyruvate and leading to ncBCAA caused by the expected intracellular accumulation of pyruvate and
de-regulation of leu under such cultivation conditions, so that higher leuA expression would then not
trigger more carbon flux to ncBCAA formation.
Up-regulation of ilvC was hypothesized to trigger reduction of ncBCAA biosynthesis, since that would
stimulate metabolic flux from α-ketobutyrate through the isoleucine biosynthetic pathway, thereby
relatively reducing α-ketobutyrate disponibility for ncBCAA formation. Results confirmed hypothesis
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
138
Discussion
since, for ilvC-tunable E. coli under standard cultivation conditions, norvaline and norleucine
concentrations present in both tested protein fractions progressively decreased by adding increasing
concentrations of L-arabinose into the medium, i.e. by increasing ilvC expression. The same trend
was observed when pyruvate pulsing and O2 limitation conditions were applied. As expected, under
those cultivation conditions, ncBCAA concentrations were higher.
AHAS III has more substrate preference for α-ketobutyrate than pyruvate (Barak et al., 1987; Salmon
et al., 2006; Vinogradow et al., 2006). Hence, an increase in AHAS III concentration would favor the
metabolic flux from α-ketobutyrate into α-acetohydroxy-butyrate at the expense of the alternative
enzymatic reaction leading to ncBCAA synthesis. Results confirmed this hypothesis since for ilvIH-
tunable E. coli under standard cultivation conditions, norvaline and norleucine concentrations
present in both tested protein fractions progressively decreased by adding increasing concentrations
of L-arabinose into the medium, i. e. by increasing ilvIH expression. Surprisingly, under cultivation
conditions subjected to pyruvate pulses and O2 limitation, the opposite trend was observed:
increasing ilvIH expression was translated into an increase of ncBCAA concentrations. The
simultaneous presence of multiple factors playing a role in the regulation of the BCAA metabolic
pathway under mentioned cultivation conditions made challenging to find a plausible explanation for
those observations.
The ilvGM-tunable E. coli strain also reported a reduction of norvaline and norleucine concentrations
compared with the wild type E. coli strain under both tested cultivation conditions. As expected,
under cultivation conditions subjected to pyruvate pulsing and O2 limitation, ncBCAA concentrations
were higher. Moreover, the highest ncBCAA reduction observed among all tested mutant strains was
reported for this mutant. This might be because AHAS II shows the highest substrate preference for
α-ketobutyrate among the AHAS isozymes, being then most of the metabolic flux from α-
ketobutyrate directed to the isoleucine pathway. In addition, kcat/Km of AHAS II is about 20-fold
higher than AHAS III (Barak et al., 1987; Salmon et al., 2006; Vinogradow et al., 2006). However,
increasing of ilvGM expression did not translate into significant variation of ncBCAA concentrations.
That might be explained by a negative feedback regulation due to an excess of free isoleucine, which
would decrease AHAS II activity (Salmon et al., 2006), thus counteracting the increased AHAS II
activity driven by L-arabinose addition. Thus, in the ideal case, ilvGM expression should be up-
regulated as much as possible but without leading to an accumulation of isoleucine.
Unexpectedly, the ilvBN-tunable E. coli strain reported the opposite behavior than the tested ilvIH-
and ilvGM-tunable E. coli strains under standard cultivation conditions: norvaline and norleucine
concentrations progressively increased by adding increasing concentrations of L-arabinose into the
medium, i. e. by increasing ilvBN expression. As opposed to AHAS II and III, AHAS I prefers pyruvate
than α-ketobutyrate as substrate (Barak et al., 1987; Salmon et al., 2006; Vinogradow et al., 2006).
Hence, an increase in AHAS I concentration would favor the metabolic flux from pyruvate into the
valine and leucine biosynthetic pathway at the expense of pyruvate transformation to α-
ketobutyrate by the leu operon. However, up-regulation of ilvBN might trigger overproduction of
valine, which is demonstrated to inhibit enzymatic activity of AHAS III by feedback regulation and to
activate enzymatic activity of L-threonine dehydratase (ilvA-encoded enzyme) (Salmon et al., 2006).
According to this configuration, more α-ketobutyrate would be produced from the threonine
pathway and, since activity of AHAS III is feedback regulated, α-ketobutyrate might then preferably
enter the ncBCAA biosynthetic pathway to the detriment of the isoleucine biosynthetic pathway. This
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Discussion
139
metabolic configuration might explain why, as opposed to first hypothesized, a down-regulation of
operon ilvBN triggers reduction of ncBCAA biosynthesis. The same trend was observed when
pyruvate pulsing and O2 limitation conditions were applied but, interestingly, ncBCAA concentrations
were lower than under standard conditions. The simultaneous presence of multiple factors playing a
role in the regulation of the BCAA metabolic pathway under mentioned cultivation conditions made
challenging to find a plausible explanation for those observations.
The thrA-tunable E. coli strain showed a reduction of norvaline and norleucine concentrations in both
tested protein fractions compared with the wild type strain under both tested cultivation conditions.
As expected, under cultivation conditions subjected to pyruvate pulsing and O2 limitation, ncBCAA
concentrations were higher. However, effect of increasing L-arabinose concentrations didnot show a
clear effect on ncBCAA concentrations. With 0.4% L-ara, a significant decrease in norvaline and
norleucine was observed. However, a further increase in thrA expression when using higher L-ara
concentrations did not show any further variation of ncBCAA concentration if compared with 0.4% L-
ara. That might be explained by a bottleneck taking place in the metabolic pathway downstream of
thrA, so that a higher thrA expression does not translate in higher α-ketobutyrate production. In
addition, it is well known that L-threonine allosterically inhibits enzyme activity of thrA-encoded
enzyme so that an excessive L-ara induction triggers accumulation of threonine, then causing
feedback inhibition of enzymatic activity, hence annulling the higher expected thrA expression. This
observation might indicate that the threonine metabolic pathway is not the one redirecting more flux
to α-ketobutyrate but the pyruvate pathway.
The ilvA-tunable E. coli strain showed that, mainly in the intracellular protein soluble protein fraction,
norvaline and norleucine concentrations report a reduction compared with control strain under both
tested cultivation conditions but effect of increasing L-arabinose concentrations did not show a clear
trend on variation of ncBCAA concentrations. As expected, under cultivation conditions subjected to
pyruvate pulsing and O2 limitation, ncBCAA concentrations were higher. With 0.05% L-ara, a
significant decrease in norvaline and norleucine levels was observed. However, a further increase in
ilvA expression when using higher L-ara concentrations did not show any further reduction of
ncBCAA concentration if compared with 0.05% L-ara. As for thrA, that might be explained by a
bottleneck in the metabolic pathway downstream of ilvA. Moreover, it is well known that isoleucine
allosterically inhibits enzyme activity of ilvA-encoded enzyme. May then be that an excessive L-ara
induction triggers accumulation of isoleucine, then causing feedback inhibition of enzymatic activity,
hence annulling the higher expected ilvA expression. As stated before, this observation might
indicate that the threonine metabolic pathway is not the one redirecting more flux to α-ketobutyrate
but the pyruvate pathway.
Unlike norvaline and norleucine, concentration of β-methylnorleucine did not significantly change
after modulating expression of target genes. This suggests that the metabolic effect triggered by
tuning expression of target genes involved in the BCAA biosynthetic pathway has a limited influence
on β-methylnorleucine synthesis, in contrast to norleucine and norvaline. As aforementioned, the
fact that β-methylnorleucine is the ncBCAA, the biosynthetic pathway of which is the longest starting
from pyruvate, also supports this hypothesis. Norvaline and norleucine biosynthesis require the
enzymes of the leu operon to transform pyruvate to the ncBCAA precursors α-ketovalerate and α-
ketocaproate while β-methylnorleucine biosynthesis demands the additional actuation of the
enzymes of the ilv operon in order to convert α-ketovalerate into the β-methylnorleucine precursor
α-keto-β-metylcaproate. Observed results might as well suggest that β-methylnorleucine is
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
140
Discussion
synthesized by an independent alternative metabolic pathway. However, no references were found
in the literature supporting this hypothesis.
According to the previous discussion, research hypothesis 1 was partially accepted.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Conclusions and Outlook
141
7. Conclusions and Outlook
This thesis was aimed to close the scientific gap by developing a new scientific approach in order to
reduce ncBCAA biosynthesis and subsequent mis-incorporation into recombinant proteins since all
the alternative strategies published so far present numerous disadvantages which challenge their
effective application. In this study we demonstrated that ncBCAA biosynthesis can be exogenously
controlled by fine tuning expression of target genes involved in the BCAA biosynthetic pathway. By
screening different engineered tunable E. coli mutants in a mini-reactor system we demonstrated
that an up-regulation of ilvC, ilvIH and ilvGM and down-regulation of leuA and ilvBN trigger a
reduction of norvaline and norleucine biosynthesis and mis-incorporation into recombinant mini-
proinsulin expressed in this study. Concerning target genes ilvA and thrA, results suggest that the
threonine pathway is not the one redirecting more metabolic flux to α-ketobutyrate. Among the
tested genes, up-regulation of ilvIH and ilvGM showed the highest reduction of ncBCAA biosynthesis
and mis-incorporation. Strain screening was performed in fed-batch mode under standard cultivation
conditions as well as under cultivation conditions subjected to pyruvate pulsing and O2 limitation.
The latter cultivation strategy was demonstrated to mimic large-scale effects since biosynthesis of
ncBCAA and metabolites of the overflow metabolism were reported under those conditions. This
cultivation approach represents a novel scale-down strategy which could contribute to accelerate the
process of strain screening. Interestingly, norleucine was the most mis-incorporated ncBCAA and β-
methylnorleucine levels did not significantly change under tested experimental conditions, which
may suggest that β-methylnorleucine is synthesized by an alternative unknown metabolic pathway.
Potential ilvIH and ilvGM-tunable E. coli strains showing a preferred protein impurity profile during
screening in the mini-reactor system were further verified in a 15L reactor under cultivation
conditions subjected to pyruvate pulsing and O2 limitation, reporting same recombinant mini-
proinsulin production and a highly significant reduction of ncBCAA mis-incorporation into the
recombinant protein in comparison with the non-engineered E. coli strain. This is in accordance with
what was reported at mini-reactor level, thus confirming the reliability and robustness of the
aforementioned strains. These novel E. coli strains might then be employed as expression hosts in
large-scale reactors for industrial production of recombinant proteins with a reduced ncBCAA mis-
incorporation profile. Testing of the engineered E. coli strains in alternative scale-down reactors such
as a 2-compartment scale-down reactor (Limberg et al., 2016) would give more hints about their
applicability in industrial scale recombinant protein production processes.
In order to ensure optimal expression of the target gene it was necessary to expose the
corresponding engineered tunable E. coli strain to a certain L-arabinose concentration and this was
only possible by exogenously adding the inducer molecule into the system, which is not optimal from
an economic and operational point of view if recombinant protein production is intended in
industrial scale. The ideal scenario would be that the optimal expression level of a target gene
resulting in a reduced ncBCAA mis-incorporation would be maintained constant during the E. coli
fermentation process without the need of adding external chemical compounds. A number of genetic
engineering strategies such as promoter engineering and antisense RNA technology could be applied
in order to achieve that preferred scenario. The development of novel endogenous constitutive
promoters through rational design or random mutagenesis allowing optimal expression levels of a
target gene would allow production of recombinant proteins with a reduced ncBCAA mis-
incorporation profile without the need of transforming additional plasmids into the E. coli expression
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
142 Conclusions and Outlook
host. This approach might be appropriate for both genetic up- and down-regulation purposes. An
antisense RNA (asRNA) is a single stranded RNA complementary to a certain mRNA so that both RNA
molecules can hybridize, blocking mRNA translation. This strategy might be useful for genetic down-
regulation purposes. However, additional transformation of a plasmid enabling asRNA expression
might be necessary in this case.
Moreover, optimal expression levels of single target genes leading to a reduction of ncBCAA mis-
incorporation into recombinant mini-proinsulin were investigated in this current thesis. However, it
might be also very interesting to investigate expression regulation of two or more target genes
simultaneously in order to achieve further reduction of ncBCAA mis-incorporation that might not be
possible by just considering a single gene. This might be possible by using alternative tunable
promoters, one different for each target gene. For instance, in order to evaluate effect of expression
regulation of two different target genes simultaneously in an E. coli host, an arabinose-inducible
promoter could be used for one gene while a XylSPm promoter could be employed for the other.
Taking into consideration results obtained in this investigation, the following two genetic
combinations are predicted to report the most improved recombinant protein impurity profile: (i)
down-regulation of gene leuA and repair or up-regulation of operon ilvGM, (ii) down-regulation of
gene leuA and up-regulation of operon ilvIH. The use of alternative molecular methods such as
promoter engineering and antisense RNA technology might also be considered for this purpose.
In order to better understand how endogenous genetic regulation affects ncBCAA biosynthesis it
would be interesting to analyze the content of free amino acids present in the cytosol. In this study
the amino acid content was analyzed in the inclusion body fraction and in the intracellular protein
soluble fraction (including the free amino acids). Since amino acid depletion or amino acid excess
triggers several genetic regulation mechanisms affecting expression of genes involved in the BCAA
biosynthetic pathway at both transcriptional (e. g. attenuation, stringent response) and
posttranslational level (e. g. feedback inhibition) that could modulate ncBCAA biosynthesis, such
additional data would be valuable to complete the discussion of some of the results generated in this
thesis. Furthermore, the experimental design and analysis carried out in this thesis were considered
sufficient to determine if an up- or down-regulation of a certain target gene causes a reduction of
ncBCAA biosynthesis, by measuring ncBCAA levels under different inducer concentrations. However,
it would be interesting to analyze the actual expression levels of each tested target gene under
different L-arabinose concentrations by means of transcriptomics or proteomics technologies in
order to identify the optimal expression level of a target gene triggering the most reduction of
ncBCAA biosynthesis. In a most preferred scenario, the availability of metabolic flux data would
contribute to obtain a clear picture of the metabolic stand of the BCAA biosynthetic pathway
resulting from each experimental scenario.
Further strategies aiming reduction of ncBCAA mis-incorporation, which are not yet available in the
state of the art and that, together with the approach investigated in this thesis, might as well be
considered as potential comprise: (i) substitution of codons reporting higher ncBCAA mis-
incorporation from the coding region of a gene of interest, (ii) protein engineering to increase
specificity of biosynthetic enzymes for certain α-keto acids, (iii) protein engineering to increase
specificity of aminoacyl tRNA synthetases for the canonical amino acids, and (iv) repair of gene ilvG in
the genome of E. coli K-12 strains to recover AHAS II activity.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
143
8. References
Abu-Absi, N., Inlow, D., Macdonald, A. C., Poulhzan, M. (2008). Preventing norvaline and norleucine
mis-incorporation in recombinant proteins. US Patent Application WO 2007/103521 A3.
Afroz, T., Biliouris, K., Kaznessis, Y., & Beisel, C. L. (2014). Bacterial sugar utilization gives rise to
distinct single‐cell behaviours. Molecular microbiology,93(6), 1093-1103.
Ahmadi, M. K., Fang, L., Moscatello, N., & Pfeifer, B. A. (2016). E. coli metabolic engineering for gram
scale production of a plant-based anti-inflammatory agent. Metabolic Engineering, 38, 382-
388.
Alba, B. M. & Gross, C. A. (2004). Regulation of the Escherichia coli sigma-dependent envelope stress
response. Mol Microbiol 52, 613–619.
Alvarez-Carreño, Claudia, Arturo Becerra, and Antonio Lazcano. "Norvaline and Norleucine may have
been more abundant protein components during early stages of cell evolution." Origins of
Life and Evolution of Biospheres 43.4-5 (2013): 363-375.
Amann, E., Brosius, J., & Ptashne, M. (1983). Vectors bearing a hybrid trp-lac promoter useful for
regulated expression of cloned genes in Escherichia coli.Gene, 25(2), 167-178.
Anane E, Lòpez CDC, Neubauer P, Cruz Bournazou MN (2017) Modelling overflow metabolism in
Escherichia coli by acetate cycling. Biochem Engin J 125(15):23–30
Anane, E., Sawatzki, A., Neubauer, P., & Cruz‐Bournazou, M. N. (2019). Modelling concentration
gradients in fed‐batch cultivations of E. coli–towards the flexible design of scale‐down
experiments. Journal of Chemical Technology & Biotechnology, 94(2), 516-526.
Andersen DC, Swartz J; Ryll T, Lin N, Snedecor B. 2001. Metabolic Oscillations in an E. coli
Fermentation. Biotechnol. Bioen. 75(2): 212-218.
Apostol, I., Levine, J., Lippincott, J., Leach, J., Hess, E., Glascock, C. B., ... & Blackmore, R. (1997).
Incorporation of norvaline at leucine positions in recombinant human hemoglobin expressed
in Escherichia coli. Journal of Biological Chemistry, 272(46), 28980-28988.
Arditti, R., Grodzicker, T., & Beckwith, J. (1973). Cyclic adenosine monophosphate-independent
mutants of the lactose operon of Escherichia coli.Journal of bacteriology, 114(2), 652-655.
Arfin, S.M., Ratzkin, B. & Umbarger, H.E. (1969) The metabolism of valine and isoleucine
in Escherichia coli. XVII. The role of induction in the derepression of acetohydroxy acid
isomeroreductase. Biochem Biophys Res Commun 37: 902–908.
Aris, A., Corchero, J. L., Benito, A., Carbonell, X., Viaplana, E., & Villaverde, A. (1998). The expression
of recombinant genes from bacteriophage lambda strong promoters triggers the SOS
response in Escherichia coli. Biotechnology and bioengineering, 60(5), 551-559.
Artsimovitch, I., Patlan, V., Sekine, S. I., Vassylyeva, M. N., Hosaka, T., Ochi, K., ... & Vassylyev, D. G.
(2004). Structural basis for transcription regulation by alarmone ppGpp. Cell, 117(3), 299-
310.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
144
References
Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., ... & Mori, H. (2006). Construction of
Escherichia coli K‐12 in‐frame, single‐gene knockout mutants: the Keio collection. Molecular
systems biology, 2(1).
Baccigalupi, L., Marasco, R., Ricca, E., De Felice, M., & Sacco, M. (1995). Control of ilvIH transcription
during amino acid downshift in stringent and relaxed strains of Escherichia coli. FEMS
microbiology letters, 131(1), 95-98.
Bailey, J. E. (1991, June). Toward a science of metabolic engineering. American Association for the
Advancement of Science.
Balzer, S., Kucharova, V., Megerle, J., Lale, R., Brautaset, T., & Valla, S. (2013). A comparative analysis
of the properties of regulated promoter systems commonly used for recombinant gene
expression in Escherichia coli. Microbial cell factories, 12(1), 1.
Barak, Z., Chipman, D. M., & Gollop, N. A. T. A. N. (1987). Physiological implications of the specificity
of acetohydroxy acid synthase isozymes of enteric bacteria. Journal of bacteriology, 169(8),
3750-3756);
Basturea, G. N., Zundel, M. A., & Deutscher, M. P. (2011). Degradation of ribosomal RNA during
starvation: comparison to quality control during steady-state growth and a role for RNase PH.
Rna, 17(2), 338-345.
Beals JM, DeFelippis MR, Kovach PM. Chaper 12: Insulin. In: Crommelin DJA, Sindelar RD, Meibohm B
(eds): Pharmaceutical biotechnology: fundamentals and application. 3rd ed. 12:265-280,
Informa Healthcare (2008).
Biermann, M., Linnemann, J., Knüpfer, U., Vollstädt, S., Bardl, B., Seidel, G., & Horn, U. (2013). Trace
element associated reduction of norleucine and norvaline accumulation during oxygen
limitation in a recombinant Escherichia coli fermentation. Microbial cell factories, 12(1), 1.
Binder, D., Bier, C., Grünberger, A., Drobietz, D., Hage‐Hülsmann, J., Wandrey, G., ... & Jaeger, K. E.
(2016). Photocaged Arabinose: A Novel Optogenetic Switch for Rapid and Gradual Control of
Microbial Gene Expression. ChemBioChem, 17(4), 296-299.
Binder, D., Grünberger, A., Loeschcke, A., Probst, C., Bier, C., Pietruszka, J., ... & Drepper, T. (2014).
Light-responsive control of bacterial gene expression: precise triggering of the lac promoter
activity using photocaged IPTG.Integrative biology, 6(8), 755-765.
Biryukova, I. V., Krylov, A. A., Kiseleva, E. M., Minaeva, N. I., & Mashko, S. V. (2010). Construction of
the new Escherichia coli K12 MG 1655 novel strain with improved growth characteristics for
application in metabolic engineering. Russian journal of genetics, 46(3), 308-314.
Blatny, J. M., Brautaset, T., Winther-Larsen, H. C., Karunakaran, P., & Valla, S. (1997). Improved
broad-host-range RK2 vectors useful for high and low regulated gene expression levels in
gram-negative bacteria. Plasmid, 38(1), 35-51.
Blattner, F. R., Plunkett, G., Bloch, C. A., Perna, N. T., Burland, V., Riley, M., ... & Gregor, J. (1997). The
complete genome sequence of Escherichia coli K12. Science, 277(5331), 1453-1462.
Bloor, A. E., & Cranenburgh, R. M. (2006). An efficient method of selectable marker gene excision by
Xer recombination for gene replacement in bacterial chromosomes. Applied and
environmental microbiology, 72(4), 2520-2525.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
145
Bogosian, G., O’Neil, J. P., Smith, H. Q. (2013). Prevention of incorporation of non-standard amino
acids into proteins. US Patent 8603781.
Bogosian, G., Violand, B. N., Dorward-King, E. J., Workman, W. E., Jung, P. E., & Kane, J. F. (1989).
Biosynthesis and incorporation into protein of norleucine by Escherichia coli. Journal of
Biological Chemistry, 264(1), 531-539.
Bowden, K. E., Wiese, N. S., Perwez, T., Mohanty, B. K., & Kushner, S. R. (2017). The rph-1 encoded
truncated RNase PH protein inhibits RNase P maturation of pre-tRNAs with short leader
sequences in the absence of RppH. Journal of bacteriology, JB-00301.
Brautaset, T., Lale, R., & Valla, S. (2009). Positively regulated bacterial expression systems. Microbial
biotechnology, 2(1), 15-30.
Brautaset, T., Petersen, S. B., & Valla, S. (2000). In vitro determined kinetic properties of mutant
phosphoglucomutases and their effects on sugar catabolism in Escherichia coli. Metabolic
engineering, 2(2), 104-114.
Brown, T. D. K., Jones-Mortimer, M. C., & Kornberg, H. L. (1977). The enzymic interconversion of
acetate and acetyl-coenzyme A in Escherichia coli. Microbiology, 102(2), 327-336.
Brückner, R., & Titgemeyer, F. (2002). Carbon catabolite repression in bacteria: choice of the carbon
source and autoregulatory limitation of sugar utilization. FEMS microbiology letters, 209(2),
141-148.
Brunner et al. (1997). Fermentation media and methods for controlling norleucine in polypeptides.
US Pat. No. 5698418
Busby, S., & Ebright, R. H. (1999). Transcription activation by catabolite activator protein
(CAP). Journal of molecular biology, 293(2), 199-213.
Bylund, F., Collet, E., Enfors, S. O., & Larsson, G. (1998). Substrate gradient formation in the large-
scale bioreactor lowers cell yield and increases by-product formation. Bioprocess
Engineering, 18(3), 171-180.
Calos, M. P. (1978). DNA sequence for a low-level promoter of the lac repressor gene and
anup'promoter mutation. Nature, 274, 762-765.
Calos, M. P., & Miller, J. H. (1980). DNA sequence alteration resulting from a mutation impairing
promoter function in the lac repressor gene. Molecular and General Genetics MGG, 178(1),
225-227
Calos, M. P., & Miller, J. H. (1981). The DNA sequence change resulting from the I Q1mutation, which
greatly increases promoter strength. Molecular and General Genetics MGG, 183(3), 559-560.
Campos‐Bermudez, V. A., Bologna, F. P., Andreo, C. S., & Drincovich, M. F. (2010). Functional
dissection of Escherichia coli phosphotransacetylase structural domains and analysis of key
compounds involved in activity regulation. The FEBS journal, 277(8), 1957-1966.
Cardona, S. T., & Valvano, M. A. (2005). An expression vector containing a rhamnose-inducible
promoter provides tightly regulated gene expression in Burkholderia cenocepacia. Plasmid,
54(3), 219-228.
Carneiro, S., Ferreira, E. C., & Rocha, I. (2013). Metabolic responses to recombinant bioprocesses in
Escherichia coli. Journal of biotechnology, 164(3), 396-408
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
146
References
Carrier, T. A., & Keasling, J. D. (1999). Investigating autocatalytic gene expression systems through
mechanistic modeling. Journal of theoretical biology, 201(1), 25-36.
Chen, C. F., Lan, J., Korovine, M., Shao, Z. Q., Tao, L., Zhang, J., & Newman, E. B. (1997). Metabolic
regulation of Irp gene expression in Escherichia coli K12. Microbiology, 143(6), 2079-2084.
Chen, S., Iannolo, M., & Calvo, J. M. (2005). Cooperative binding of the leucine-responsive regulatory
protein (Lrp) to DNA. Journal of molecular biology, 345(2), 251-264.
Chen, X., Ma, J. J., Tan, M., Yao, P., Hu, Q. H., Eriani, G., & Wang, E. D. (2011). Modular pathways for
editing non-cognate amino acids by human cytoplasmic leucyl-tRNA synthetase. Nucleic acids
research, 39(1), 235-247.
Chen, X., Zaro, J. L., & Shen, W. C. (2013). Fusion protein linkers: property, design and functionality.
Advanced drug delivery reviews, 65(10), 1357-1369.
Chen, Z., Sun, X., Li, Y., Yan, Y., & Yuan, Q. (2017). Metabolic engineering of Escherichia coli for
microbial synthesis of monolignols. Metabolic Engineering.
Choi, J. H., Keum, K. C., & Lee, S. Y. (2006). Production of recombinant proteins by high cell density
culture of Escherichia coli. Chemical Engineering Science, 61(3), 876-885.
Cirino, P. C., Tang, Y., Takahashi, K., Tirrell, D. A., & Arnold, F. H. (2003). Global incorporation of
norleucine in place of methionine in cytochrome P 450 BM-3 heme domain increases
peroxygenase activity. Biotechnology and bioengineering, 83(6), 729-734.
Cohen, G. N., & Munier, R. (1959). Effets des analogues structuraux d'aminioacides sur la croissance.
la synthese de proteines et la synthese d'enzymes chez Escherichia coli. Biochem. Biophys.
Acta 31(2): 347-356).
Cohen, S. N., Chang, A. C., & Hsu, L. (1972). Nonchromosomal antibiotic resistance in bacteria:
genetic transformation of Escherichia coli by R-factor DNA. Proceedings of the National
Academy of Sciences, 69(8), 2110-2114.
Court, D. L., Sawitzke, J. A., & Thomason, L. C. (2002). Genetic engineering using homologous
recombination. Annual review of genetics, 36(1), 361-388.
Couturier, E., & Rocha, E. P. (2006). Replication‐associated gene dosage effects shape the genomes of
fast‐growing bacteria but only for transcription and translation genes. Molecular
microbiology, 59(5), 1506-1518.
Cowie, D. B., Cohen, G. N., Bolton, E. T., & De Robichon-Szulmajster, H. (1959). Amino acid analog
incorporation into bacterial proteins. Biochimica et biophysica acta, 34, 39-46.
Cvak, L., Jegorov, A., Sedmera, P., Císařová, I., Čejka, J., Kratochvíl, B., & Pakhomova, S. (2005).
Norleucine, a natural occurrence in a novel ergot alkaloid γ-ergokryptinine. Amino
acids, 29(2), 145-150.
Cvetešić, N., Akmačić, I., & Gruić-Sovulj, I. (2013). Lack of discrimination against non-proteinogenic
amino acid norvaline by elongation factor Tu from Escherichia coli. Croatica Chemica
Acta, 86(1), 73-82.
Cvetesic, N., Palencia, A., Halasz, I., Cusack, S., & Gruic‐Sovulj, I. (2014). The physiological target for
LeuRS translational quality control is norvaline. The EMBO journal, e201488199.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
147
Cvetesic, N., Perona, J. J., & Gruic-Sovulj, I. (2012). Kinetic partitioning between synthetic and editing
pathways in class I aminoacyl-tRNA synthetases occurs at both pre-transfer and post-transfer
hydrolytic steps. Journal of Biological Chemistry, 287(30), 25381-25394.
Cvetesic, N., Semanjski, M., Soufi, B., Krug, K., Gruic-Sovulj, I., & Macek, B. (2016). Proteome-wide
measurement of non-canonical bacterial mistranslation by quantitative mass spectrometry of
protein modifications. Scientific reports, 6, 28631.
Daber, R., Stayrook, S., Rosenberg, A., & Lewis, M. (2007). Structural analysis of lac repressor bound
to allosteric effectors. Journal of molecular biology,370(4), 609-619.
Dailey, F. E., & Cronan, J. E. (1986). Acetohydroxy acid synthase I, a required enzyme for isoleucine
and valine biosynthesis in Escherichia coli K12 during growth on acetate as the sole carbon
source. Journal of bacteriology, 165(2), 453-460.
Datsenko, K. A., & Wanner, B. L. (2000). One-step inactivation of chromosomal genes in Escherichia
coli K12 using PCR products. Proceedings of the National Academy of Sciences, 97(12), 6640-
6645.
Davidson, I. (1997). Hydrolysis of samples for amino acid analysis. In Protein Sequencing
Protocols (pp. 119-129). Humana Press, Totowa, NJ.
De Boer, H. A., Comstock, L. J., & Vasser, M. (1983). The tac promoter: a functional hybrid derived
from the trp and lac promoters. Proceedings of the National Academy of Sciences, 80(1), 21-
25.
de los Rios, S., & Perona, J. J. (2007). Structure of the Escherichia coli leucine-responsive regulatory
protein Lrp reveals a novel octameric assembly. Journal of molecular biology, 366(5), 1589-
1602.
DeFelice, M., and M. Levinthal. 1977. The acetohydroxy acid synthase III isoenzyme of Escherichia
coli K12: regulation of synthesis by leucine. Biochem. Biophys. Res. Commun. 79:82–87
Dong, H., Nilsson, L., & Kurland, C. G. (1996). Co-variation of trna abundance and codon usage
inescherichia coliat different growth rates. Journal of molecular biology, 260(5), 649-663.
Dong, X., Chen, X., Qian, Y., Wang, Y., Wang, L., Qiao, W., & Liu, L. (2016). Metabolic engineering of
Escherichia coli W3110 to produce L‐malate. Biotechnology and Bioengineering.
Dörschung, P.; Habermann, P.; Seipke, G.; Uhlmann, E. (1989). Mini-Proinsulin, seine Herstellung und
Verwendung. Patent number: EP0347781B1.
Egan, S. M., & Schleif, R. F. (1993). A regulatory cascade in the induction of rhaBAD. Journal of
molecular biology, 234(1), 87-98.
Eiteman, M. A., & Altman, E. (2006). Overcoming acetate in Escherichia coli recombinant protein
fermentations. Trends in biotechnology, 24(11), 530-536.
Enfors, S. O., Jahic, M., Rozkov, A., Xu, B., Hecker, M., Jürgen, B., ... & Noisommit-Rizzi, N. (2001).
Physiological responses to mixing in large-scale bioreactors. Journal of biotechnology, 85(2),
175-185
Fahnert, B., Lilie, H., & Neubauer, P. (2004). Inclusion bodies: formation and utilisation.
In Physiological Stress Responses in Bioprocesses (pp. 93-142). Springer, Berlin, Heidelberg.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
148
References
Fang, M., & Bauer, C. E. (2018). Regulation of stringent factor by branched-chain amino acids.
Proceedings of the National Academy of Sciences, 201803220.
Fenton, D., Lai, P. H., Lu, H., Mann, M., Tsai, L. (1994). Control of norleucine incorporation into
recombinant proteins. US Patent 5599690.
Ferrer-Miralles, N., Domingo-Espín, J., Corchero, J. L., Vázquez, E., & Villaverde, A. (2009). Microbial
factories for recombinant pharmaceuticals. Microbial cell factories, 8(1), 1.
Fiedler, S., & Wirth, R. (1988). Transformation of bacteria with plasmid DNA by electroporation.
Analytical biochemistry, 170(1), 38-44.
Freundlich, M. (1977). Cyclic AMP can replace the relA-dependent requirement for derepression of
acetohydroxy acid synthase in E. coli K12. Cell, 12(4), 1121-1126.
Freundlich, M., Ramani, N., Mathew, E., Sirko, A., & Tsui, P. (1992). The role of integration host factor
in gene expression in Escherichia coli. Molecular microbiology, 6(18), 2557-2563.
Friden, P., Newman, T., & Freundlich, M. (1982). Nucleotide sequence of the ilvB promoter-
regulatory region: a biosynthetic operon controlled by attenuation and cyclic AMP.
Proceedings of the National Academy of Sciences, 79(20), 6156-6160.
Friden, P., Tsui, P., Okamoto, K., & Freundlich, M. (1984). Interaction of cyclic AMP receptor protein
with the ilvB biosynthetic operon in E. coli. Nucleic acids research, 12(21), 8145-8160. (I)
Friden, P., Voelkel, K., Sternglanz, R., & Freundlich, M. (1984). Reduced expression of the isoleucine
and valine enzymes in integration host factor mutants of Escherichia coli. Journal of
molecular biology, 172(4), 573-579. (II)
Friedman, D. I. (1988). Integration host factor: a protein for all reasons. Cell, 55(4), 545-554.
García-Fruitós, E. (2010). Inclusion bodies: a new concept. Microbial cell factories, 9(1), 80.
Giacalone. (2006). Toxic protein expression in Escherichia coli using a rhamnose-based tightly
regulated and tunable promoter system (vol 40, pg 355, 2006). Biotechniques, 40(5), 596-
596.
Gill, R. T., Valdes, J. J., & Bentley, W. E. (2000). A comparative study of global stress gene regulation
in response to overexpression of recombinant proteins in Escherichia coli. Metabolic
engineering, 2(3), 178-189.
Gilles, A. M., Marliere, P., Rose, T., Sarfati, R., Longin, R., Meier, A., ... & Barzu, O. (1988).
Conservative replacement of methionine by norleucine in Escherichia coli adenylate kinase.
Journal of Biological Chemistry, 263(17), 8204-8209.
Glascock, C. B., & Weickert, M. J. (1998). Using chromosomal lacIQ1 to control expression of genes
on high-copy-number plasmids in Escherichia coli.Gene, 223(1), 221-231.
Glick, B. R. (1995). Metabolic load and heterologous gene expression. Biotechnology advances, 13(2),
247-261.
Gollop, N., Damri, B., Barak, Z. E., & Chipman, D. M. (1989). Kinetics and mechanism of acetohydroxy
acid synthase isozyme III from Escherichia coli. Biochemistry, 28(15), 6310-6317.
Görke, B., & Stülke, J. (2008). Carbon catabolite repression in bacteria: many ways to make the most
out of nutrients. Nature Reviews Microbiology, 6(8), 613-624.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
149
Grenier, F., Matteau, D., Baby, V., & Rodrigue, S. (2014). Complete genome sequence of Escherichia
coli BW25113. Genome announcements, 2(5), e01038-14.
Guzman, L. M., Belin, D., Carson, M. J., & Beckwith, J. O. N. (1995). Tight regulation, modulation, and
high-level expression by vectors containing the arabinose PBAD promoter. Journal of
bacteriology, 177(14), 4121-4130.
Habermann, P., & Wengermayer, F. (1996). Fusion protein comprising an Interleukin-2 fragment
ballast portion. Patent number: 5496924.
Habermann, Paul (1986). Fusionsproteine mit eukaryotischem Ballastanteil. Patent number:
EP0229998A2.
Hanahan, D. (1983). Studies on transformation of Escherichia coli with plasmids. Journal of molecular
biology, 166(4), 557-580.
Haqqi, T., Zhao, X., Panciu, A., & Yadav, S. (2002). Sequencing in the presence of betaine:
Improvement in sequencing of the localized repeat sequence regions. Journal of
Biomolecular Techniques: JBT, 13(4), 265.
Harris, R. P., & Kilby, P. M. (2014). Amino acid mis-incorporation in recombinant biopharmaceutical
products. Current opinion in biotechnology, 30, 45-50.
Hartley, D. L., & Kane, J. F. (1988). Properties of inclusion bodies from recombinant Escherichia coli.
Hauryliuk, V., Atkinson, G. C., Murakami, K. S., Tenson, T., & Gerdes, K. (2015). Recent functional
insights into the role of (p) ppGpp in bacterial physiology. Nature Reviews Microbiology,
13(5), 298.
Hayashi, K., Morooka, N., Yamamoto, Y., Fujita, K., Isono, K., Choi, S., ... & Horiuchi, T. (2006). Highly
accurate genome sequences of Escherichia coli K‐12 strains MG1655 and W3110. Molecular
systems biology, 2(1).
Hermsen, R., Okano, H., You, C., Werner, N., & Hwa, T. (2015). A growth‐rate composition formula for
the growth of E. coli on co‐utilized carbon substrates. Molecular systems biology, 11(4), 801.
Hirokawa, Y., Kawano, H., Tanaka-Masuda, K., Nakamura, N., Nakagawa, A., Ito, M., ... & Ogasawara,
N. (2013). Genetic manipulations restored the growth fitness of reduced-genome Escherichia
coli. Journal of bioscience and bioengineering, 116(1), 52-58.
Huang, Y. T., Lyu, S. Y., Chuang, P. H., Hsu, N. S., Li, Y. S., Chan, H. C., ... & Li, T. L. (2009). In vitro
Characterization of Enzymes Involved in the Synthesis of Nonproteinogenic Residue (2S, 3S)‐
β‐Methylphenylalanine in Glycopeptide Antibiotic Mannopeptimycin. Chembiochem, 10(15),
2480-2487.
Hunter, M. F., & Parker, E. J. (2014). Modifying the determinants of α‐ketoacid substrate selectivity in
mycobacterium tuberculosis α‐isopropylmalate synthase. FEBS letters, 588(9), 1603-1607
Imber, R., Low, R. L., & Ray, D. S. (1983). Identification of a primosome assembly site in the region of
the ori 2 replication origin of the Escherichia coli mini-F plasmid. Proceedings of the National
Academy of Sciences, 80(23), 7132-7136.
Itakura, K., Hirose, T., Crea, R., Riggs, A. D., Heyneker, H. L., Bolivar, F., & Boyer, H. W. (1977).
Expression in Escherichia coli of a chemically synthesized gene for the hormone
somatostatin. Science, 198(4321), 1056-1063.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
150
References
Jensen, K. F. (1993). The Escherichia coli K12" wild types" W3110 and MG1655 have an rph
frameshift mutation that leads to pyrimidine starvation due to low pyrE expression levels.
Journal of bacteriology, 175(11), 3401-3407.
Johnson, I. S. (1983). Human insulin from recombinant DNA technology. Science, 219(4585), 632-637.
Jones, K. L., & Keasling, J. D. (1998). Construction and characterization of F plasmid‐based expression
vectors. Biotechnology and bioengineering, 59(6), 659-665.
Jones, K. L., Kim, S. W., & Keasling, J. D. (2000). Low-copy plasmids can perform as well as or better
than high-copy plasmids for metabolic engineering of bacteria. Metabolic engineering, 2(4),
328-338.
Jürgen, B., Breitenstein, A., Urlacher, V., Büttner, K., Lin, H., Hecker, M., ... & Neubauer, P. (2010).
Quality control of inclusion bodies in Escherichia coli. Microbial cell factories, 9(1), 41.
Keasling, J. D. (1999). Gene-expression tools for the metabolic engineering of bacteria. Trends in
biotechnology, 17(11), 452-460.
Kelly, C. L., Liu, Z., Yoshihara, A., Jenkinson, S. F., Wormald, M. R., Otero, J. M., ... & Estevez, R. J.
(2016). Synthetic chemical inducers and genetic decoupling enable orthogonal control of the
rhaBAD promoter. ACS synthetic biology.
Kessler, D., and J. Knappe. 1996. “Anaerobic Dissimilation of Pyruvate.” Escherichia Coli and
Salmonella: Cellular and Molecular Biology. 1: 199–216. -> I CANNOT GET IT!
Khlebnikov, A., Datsenko, K. A., Skaug, T., Wanner, B. L., & Keasling, J. D. (2001). Homogeneous
expression of the PBAD promoter in Escherichia coli by constitutive expression of the low-
affinity high-capacity AraE transporter. Microbiology, 147(12), 3241-3247.
Khlebnikov, A., Risa, Ø., Skaug, T., Carrier, T. A., & Keasling, J. D. (2000). Regulatable arabinose-
inducible gene expression system with consistent control in all cells of a culture. Journal of
Bacteriology, 182(24), 7029-7034.
Khlebnikov, A., Skaug, T., & Keasling, J. D. (2002). Modulation of gene expression from the arabinose-
inducible araBAD promoter. Journal of industrial microbiology & biotechnology, 29(1), 34-37.
Kiefhaber, T., Rudolph, R., Kohler, H. H., & Buchner, J. (1991). Protein aggregation in vitro and in vivo:
a quantitative model of the kinetic competition between folding and aggregation. Nature
biotechnology, 9(9), 825.
Kiick, K. L., Weberskirch, R., & Tirrell, D. A. (2001). Identification of an expanded set of translationally
active methionine analogues in Escherichia coli. Febs Letters, 502(1-2), 25-30.
Kisumi, M., Sugiura, M., & Chibata, I. (1976). Biosynthesis of norvaline, norleucine, and
homoisoleucine in Serratia marcescens. Journal of biochemistry, 80(2), 333-339. (I)
Kisumi, M., SUGIURA, M., Jyoji, K. A. T. O., & CHIBATA, I. (1976). L-Norvaline and L-homoisoleucine
formation by Serratia marcescens. Journal of biochemistry, 79(5), 1021-1028. (II)
Kohlhaw, G., Leary, T. R., & Umbarger, H. E. (1969). α-Isopropylmalate Synthase from Salmonella
typhimurium PURIFICATION AND PROPERTIES. Journal of Biological Chemistry, 244(8), 2218-
2225
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
151
Komori, H., Matsunaga, F., Higuchi, Y., Ishiai, M., Wada, C., & Miki, K. (1999). Crystal structure of a
prokaryotic replication initiator protein bound to DNA at 2.6 Å resolution. The EMBO journal,
18(17), 4597-4607.
Kreth, J., Lengeler, J. W., & Jahreis, K. (2013). Characterization of pyruvate uptake in Escherichia coli
K12. PLoS One, 8(6), e67125.
Laird, M. W., & Veeravalli, K. (2013). Methods and compositions for preventing norleucine mis-
incorporation into proteins. US Patent Application WO 2014/047311 A1.
Lawther RP, Calhoun DH, Adams CW, Hauser CA, Gray J, Hatfield GW. 1981. Molecular basis of valine
resistance in Escherichia coli K12. Proc. NatL Acad. Sci. USA 78(2): 922-925.
Lee, S. K., Chou, H. H., Pfleger, B. F., Newman, J. D., Yoshikuni, Y., & Keasling, J. D. (2007). Directed
evolution of AraC for improved compatibility of arabinose-and lactose-inducible promoters.
Applied and environmental microbiology, 73(18), 5711-5715.
Lee, S. Y. (1996). High cell-density culture of Escherichia coli. Trends in biotechnology, 14(3), 98-105.
Lengeler, J. W., Drews, G., & Schlegel, H. G. (Eds.). (1999). Biology of the Prokaryotes. Georg Thieme
Verlag.
Li, H., Wang, B. S., Li, Y. R., Zhang, L., Ding, Z. Y., Gu, Z. H., & Shi, G. Y. (2017). Metabolic engineering
of Escherichia coli W3110 for the production of l-methionine. Journal of Industrial
Microbiology & Biotechnology, 1-14.
Limberg MH, Pooth V, Wiechert W, Oldiges M, 2016. Plug flow versus stirred tank reactor flow
characteristics in two-compartment scale-down bioreactor: Setup-specific influence on the
metabolic phenotype and bioprocess performance of Corynebacterium glutamicum. Eng. Life
Sci. 16, 610–619
Lin, H. Y., Hanschke, R., Nicklisch, S., Riemschneider, S., Meyer, S., Gupta, A., ... & Schwahn, C. (2001).
Cellular responses to strong overexpression of recombinant genes in Escherichia coli. In
Recombinant Protein Production with prokaryotic and eukaryotic cells. A comparative view
on host physiology (pp. 55-73). Springer, Dordrecht.
Lin, R., d'Ari, R., & Newman, E. B. (1992). Lambda placMu insertions in genes of the leucine regulon:
extension of the regulon to genes not regulated by leucine. Journal of bacteriology, 174(6),
1948-1955.
Link, A. J., Mock, M. L., & Tirrell, D. A. (2003). Non-canonical amino acids in protein engineering.
Current opinion in biotechnology, 14(6), 603-609.
Liu, H., Kang, J., Qi, Q., & Chen, G. (2011). Production of lactate in Escherichia coli by redox regulation
genetically and physiologically. Applied biochemistry and biotechnology, 164(2), 162-169.
Loveland, A. B., Bah, E., Madireddy, R., Zhang, Y., Brilot, A. F., Grigorieff, N., & Korostelev, A. A.
(2016). Ribosome• RelA structures reveal the mechanism of stringent response activation.
Elife, 5, e17029.
Lu, H. S., Tsai, L. B., Kenney, W. C., & Lai, H. (1988). Identification of unusual replacement of
methionine by norleucine in recombinant interleukin-2 produced by E. coli. Biochemical and
biophysical research communications, 156(2), 807-813.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
152
References
Madigan, M., J. Martinko, K.. Bender, D. Buckley, and D. Stahl. 2014. Brock Biology of
Microorganisms. Pearson Education Limited. doi:10.2460/ajvr.75.7.613.
Madyagol, M., Al-Alami, H., Levarski, Z., Drahovská, H., Turňa, J., & Stuchlík, S. (2011). Gene
replacement techniques for Escherichia coli genome modification. Folia microbiologica,
56(3), 253-263.
Martínez-Alonso, M., González-Montalbán, N., García-Fruitós, E., & Villaverde, A. (2009). Learning
about protein solubility from bacterial inclusion bodies. Microbial cell factories, 8(1), 4.
Martinis SA, Fox GE. Non-standard amino acid recognition by Escherichia coli leucyl-tRNA
synthetase. Nucleic Acids Symp Ser. 1997;36:125–128
Mayer, M., & Buchner, J. (2004). Refolding of inclusion body proteins. In Molecular diagnosis of
infectious diseases (pp. 239-254). Humana Press.
Megerle, J. A., Fritz, G., Gerland, U., Jung, K., & Rädler, J. O. (2008). Timing and dynamics of single cell
gene expression in the arabinose utilization system. Biophysical journal, 95(4), 2103-2115.
Miller, H. I., Kikuchi, A., Nash, H. A., Weisberg, R. A., & Friedman, D. I. (1979). Site-specific
recombination of bacteriophage λ: the role of host gene products. In Cold Spring Harbor
symposia on quantitative biology (Vol. 43, pp. 1121-1126). Cold Spring Harbor Laboratory
Press.
Miyazawa, T., Yokoyama, S., & Miyake, T. (1989). Method for producing protein containing
nonprotein amino acids. U.S. Patent No. 4,879,223.
Mnatsakanyan, N., K. Bagramyan, and A. Trchounian. 2004. “Hydrogenase 3 but Not Hydrogenase 4
Is Major in Hydrogen Gas Production by Escherichia Coli Formate Hydrogenlyase at Acidic pH
and in the Presence of External Formate.” Cell Biochemistry and Biophysics 41 (3): 357–66.
doi:10.1385/CBB:41:3:357.
Mogk, A., Kummer, E., & Bukau, B. (2015). Cooperation of Hsp70 and Hsp100 chaperone machines in
protein disaggregation. Frontiers in molecular biosciences, 2, 22.
Monod, J., Wyman, J., & Changeux, J. P. (1965). On the nature of allosteric transitions: a plausible
model. J. Mol. Biol. (1965) 12, 88-118
Morgan-Kiss, R. M., Wadler, C., & Cronan, J. E. (2002). Long-term and homogeneous regulation of the
Escherichia coli araBAD promoter by use of a lactose transporter of relaxed
specificity. Proceedings of the National Academy of Sciences, 99(11), 7373-7377.
Mori, H., Kondo, A., Ohshima, A., Ogura, T., & Hiraga, S. (1986). Structure and function of the F
plasmid genes essential for partitioning. Journal of molecular biology, 192(1), 1-15.
Munier, R., & Cohen, G. N. (1956). Incorporation d'analogues structuraux d'aminoacides dans les
protéines bactériennes. Biochimica et Biophysica Acta, 21(3), 592-593.
Muramatsu, R., Misawa, S., & Hayashi, H. (2003). Finding of an isoleucine derivative of a recombinant
protein for pharmaceutical use. Journal of pharmaceutical and biomedical analysis, 31(5),
979-987.
Muramatsu, R., Negishi, T., Mimoto, T., Miura, A., Misawa, S., & Hayashi, H. (2002). Existence of β-
methylnorleucine in recombinant hirudin produced by Escherichia coli. Journal of
biotechnology, 93(2), 131-142.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
153
Murphy, K. C. (1998). Use of bacteriophage λ recombination functions to promote gene replacement
in Escherichia coli. Journal of bacteriology, 180(8), 2063-2071.
Naider, F., Z. Bohak, and J. Yariv. (1972). Reversible alkylation of a methionine residue near the active
site of ,-galactosidase. Biochemistry 11:3202-3208.
Nakamori, S., Kobayashi, S., Nishimura, T., & Takagi, H. (1999). Mechanism of L-methionine
overproduction by Escherichia coli: the replacement of Ser-54 by Asn in the MetJ protein
causes the derepression of L-methionine biosynthetic enzymes. Applied microbiology and
biotechnology, 52(2), 179-185.
Nakashima, N., & Miyazaki, K. (2014). Bacterial cellular engineering by genome editing and gene
silencing. International journal of molecular sciences, 15(2), 2773-2793.
Nandi, P. & Sen, G.P. (1953). An antifungal substance from a strain of B. subtilis. Nature 171, 871-872
Neidhardt, F. C., & Umbarger, H. E. (1987). Chemical composition of E. coli. E. coli and S.
typhimurium: Cellular and Molecular Biology.
Ni, J., Gao, M., James, A., Yao, J., Yuan, T., Carpick, B., ... & Farrell, P. (2015). Investigation into the
mis-incorporation of norleucine into a recombinant protein vaccine candidate. Journal of
industrial microbiology & biotechnology, 42(6), 971-975.
Nielsen, J., & Keasling, J. D. (2016). Engineering Cellular Metabolism. Cell, 164(6), 1185-1197.
Ning, Y., Wu, X., Zhang, C., Xu, Q., Chen, N., & Xie, X. (2016). Pathway construction and metabolic
engineering for fermentative production of ectoine in Escherichia coli. Metabolic
engineering, 36, 10-18.
Novick, A., & Weiner, M. (1957). Enzyme induction as an all-or-none phenomenon. Proceedings of
the National Academy of Sciences of the United States of America, 43(7), 553.
Olins, P. O., & Rangwala, S. H. (1989). A novel sequence element derived from bacteriophage T7
mRNA acts as an enhancer of translation of the lacZ gene in Escherichia coli. Journal of
Biological Chemistry, 264(29), 16973-16976.
Ozbudak, E. M., Thattai, M., Lim, H. N., Shraiman, B. I., & Van Oudenaarden, A. (2004). Multistability
in the lactose utilization network of Escherichia coli. Nature, 427(6976), 737-740.
Packeiser, H., Lim, C., Balagurunathan, B., Wu, J., & Zhao, H. (2013). An extremely simple and
effective colony PCR procedure for bacteria, yeasts, and microalgae. Applied biochemistry
and biotechnology, 169(2), 695-700.
Parekh, B. S., & Hatfield, G. W. (1997). Growth rate-related regulation of the ilvGMEDA operon of
Escherichia coli K12 is a consequence of the polar frameshift mutation in the ilvG gene of this
strain. Journal of bacteriology, 179(6), 2086-2088.
Parekh, B. S., Sheridan, S. D., & Hatfield, G. W. (1996). Effects of integration host factor and DNA
supercoiling on transcription from the ilvPG promoter of Escherichia coli. Journal of Biological
Chemistry, 271(34), 20258-20264.
Patte JC (1996) Biosynthesis of threonine and lysine. In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin
ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds)
Escherichia coli and Salmonella: cellular and molecular biology, 2nd edn. ASM, Washington,
pp 528–541
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
154
References
Pei, J., Dong, P., Wu, T., Zhao, L., Fang, X., Cao, F., ... & Yue, Y. (2016). Metabolic Engineering of
Escherichia coli for Astragalin Biosynthesis. Journal of Agricultural and Food Chemistry,
64(42), 7966-7972.
Phue, J. N., Noronha, S. B., Hattacharyya, R., Wolfe, A. J., & Shiloach, J. (2005). Glucose metabolism at
high density growth of E. coli B and E. coli K: differences in metabolic pathways are
responsible for efficient glucose utilization in E. coli B as determined by microarrays and
Northern blot analyses. Biotechnology and Bioengineering, 90(7), 805-820.
Pickering, M., & Newton, P. (1990). Amino acid hydrolysis: Old problems, new solutions. Lc Gc, 8(10),
778-781.
Platko, J. V., D. A. Willins, and J. M. Calvo. 1990. The ilvIH operon of Escherichia coli is positively
regulated. J. Bacteriol. 172:4563–4570.
Poulsen, P., Jensen, K. F., VALENTIN‐HANSEN, P., Carlsson, P., & Lundberg, L. G. (1983). Nucleotide
sequence of the Escherichia coli pyrE gene and of the DNA in front of the protein‐coding
region. The FEBS Journal, 135(2), 223-229.
Pribnow, D. (1975). Bacteriophage T7 early promoters: nucleotide sequences of two RNA polymerase
binding sites. Journal of molecular biology, 99(3), 419-443.
Quail, M. A., Haydon, D. J., & Guest, J. R. (1994). The pdhR–aceEF–lpd operon of Escherichia coli
expresses the pyruvate dehydrogenase complex. Molecular microbiology, 12(1), 95-104.
Ramón, A., Señorale, M., & Marín, M. (2014). Inclusion bodies: not that bad…. Frontiers in
microbiology, 5, 56.
Randhawa, Z. I., Witkowska, H. E., Cone, J., Wilkins, J. A., Hughes, P., Yamanishi, K., ... & Arthur, P.
(1994). Incorporation of norleucine at methionine positions in recombinant human
macrophage colony stimulating factor (M-CSF, 4-153) expressed in Escherichia coli: structural
analysis. Biochemistry, 33(14), 4352-4362.
Reitz C. 2011: In-vivo screening system for engineered Triosephosphate Isomerase libraries. TU
Berlin, Diploma Thesis.
Reitz, C. (2017). Impacts of oscillating cultivation conditions on the quality of recombinant inclusion
bodies in Escherichia coli. PhD Thesis.
Reitz, C., Fan, Q., & Neubauer, P. (2018). Synthesis of non-canonical branched-chain amino acids in
Escherichia coli and approaches to avoid their incorporation into recombinant proteins.
Current opinion in biotechnology, 53, 248-253.
Rhee KY, Opel M, Ito E, Hung S, Arfin SM, Hatfield GW. 1999. Transcriptional coupling between the
divergent promoters of a prototypic LysR-type regulatory system, the ilvYC operon of
Escherichia coli. Proc Natl Acad Sci USA 96:14294–14299.
Rhee, K. Y., Parekh, B. S., & Hatfield, G. W. (1996). Leucine-responsive regulatory protein-DNA
interactions in the leader region of the ilvGMEDA operon of Escherichia coli. Journal of
Biological Chemistry, 271(43), 26499-26507.
Rhee, K.Y., Senear, D.F. & Hatfield, G.W. (1998) Activation of gene expression by a ligand‐induced
conformational change of a protein–DNA complex. J Biol Chem 273: 11257–11266.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
155
Riesenberg, D. (1991). High-cell-density cultivation of Escherichia coli. Current Opinion in
Biotechnology, 2(3), 380-384.
Riesenberg, D., Menzel, K., Schulz, V., Schumann, K., Veith, G., Zuber, G., & Knorre, W. A. (1990). High
cell density fermentation of recombinant Escherichia coli expressing human interferon alpha
1. Applied microbiology and biotechnology, 34(1), 77-82.
Rinas, U., & Bailey, J. E. (1992). Protein compositional analysis of inclusion bodies produced in
recombinant Escherichia coli. Applied microbiology and biotechnology, 37(5), 609-614.
Rinas, U., Garcia-Fruitós, E., Corchero, J. L., Vázquez, E., Seras-Franzoso, J., & Villaverde, A. (2017).
Bacterial inclusion bodies: discovering their better half. Trends in biochemical sciences, 42(9),
726-737.
Roderick, S. L. (2005). The lac operon galactoside acetyltransferase. Comptes rendus
biologies, 328(6), 568-575.
Roesser, J. R., Nakamura, Y., & Yanofsky, C. (1989). Regulation of basal level expression of the
tryptophan operon of Escherichia coli. Journal of Biological Chemistry, 264(21), 12284-12288.
Salmon, K. A., Yang, C. R., & Hatfield, G. W. (2006). Biosynthesis and Regulation of the Branched-
Chain Amino Acids†. EcoSal Plus, 2(1).
Sawers, R. G., Blokesch, M., & Böck, A. (2004). Anaerobic Formate and Hydrogen Metabolism. EcoSal
Plus, 1(1).
Schleif, R. (2000). Regulation of the L-arabinose operon of Escherichia coli.Trends in Genetics, 16(12),
559-565.
Schweder, T., Krüger, E., Xu, B., Jürgen, B., Blomsten, G., Enfors, S. O., & Hecker, M. (1999).
Monitoring of genes that respond to process‐related stress in large‐scale
bioprocesses. Biotechnology and bioengineering, 65(2), 151-159.
Schweder, T., Lin, H., Jürgen, B., Breitenstein, A., Riemschneider, S., Khalameyzer, V., ... & Neubauer,
P. (2002). Role of the general stress response during strong overexpression of a heterologous
gene in Escherichia coli. Applied microbiology and biotechnology, 58(3), 330-337.
Sezonov, G., Joseleau-Petit, D., & D'Ari, R. (2007). Escherichia coli physiology in Luria-Bertani broth.
Journal of bacteriology, 189(23), 8746-8749.
Shiloach, J., & Fass, R. (2005). Growing E. coli to high cell density—a historical perspective on method
development. Biotechnology advances, 23(5), 345-357.
Siegele, D. A., & Hu, J. C. (1997). Gene expression from plasmids containing the araBAD promoter at
subsaturating inducer concentrations represents mixed populations. Proceedings of the
National Academy of Sciences, 94(15), 8168-8172.
Silverstone, A. E., Arditti, R. R., & Magasanik, B. (1970). Catabolite-insensitive revertants of lac
promoter mutants. Proceedings of the National Academy of Sciences, 66(3), 773-779.
Sirko, A., Zehelein, E., Freundlich, M., & Sawers, G. (1993). Integration host factor is required for
anaerobic pyruvate induction of pfl operon expression in Escherichia coli. Journal of
bacteriology, 175(18), 5769-5777.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
156
References
Sletta, H., Nedal, A., Aune, T. E. V., Hellebust, H., Hakvåg, S., Aune, R., ... & Brautaset, T. (2004).
Broad-host-range plasmid pJB658 can be used for industrial-level production of a secreted
host-toxic single-chain antibody fragment in Escherichia coli. Applied and environmental
microbiology, 70(12), 7033-7039.
Soini, J., Falschlehner, C., Liedert, C., Bernhardt, J., Vuoristo, J., & Neubauer, P. (2008). Norvaline is
accumulated after a down-shift of oxygen in Escherichia coli W3110. Microb. Cell Fact., 7: 1-
14. (I)
Soini, J., Ukkonen, K., & Neubauer, P. (2008). High cell density media for Escherichia coli are generally
designed for aerobic cultivations–consequences for large-scale bioprocesses and shake flask
cultures. Microbial Cell Factories, 7(1), 26. (II)
Soini, J., Ukkonen, K., & Neubauer, P. (2011). Accumulation of amino acids deriving from pyruvate in
Escherichia coli W3110 during fed-batch cultivation in a two-compartment scale-down
bioreactor. Advances in Bioscience and Biotechnology, 2(05), 336.
Soper, T. S., Doellgast, G. J., & Kohlhaw, G. B. (1976). Mechanism of feedback inhibition by leucine:
Purification and properties of a feedback-resistant α-isopropylmalate synthase. Archives of
biochemistry and biophysics, 173(1), 362-374.
Sousa, C., de Lorenzo, V., & Cebolla, A. (1997). Modulation of gene expression through chromosomal
positioning in Escherichia coli. Microbiology, 143(6), 2071-2078.
Steinsiek, S., & Bettenbrock, K. (2012). Glucose transport in Escherichia coli mutant strains with
defects in sugar transport systems. Journal of bacteriology, 194(21), 5897-5908.
Sugiura, M., Kisumi, M., & Chibata, I. (1981). Biosynthetic pathway of β-methylnorleucine, an
antimetabolite produced by Serratia marcescens. The Journal of antibiotics, 34(10), 1283-
1289.
Sunasara, K. M., Cramer, S. M., Hauer, C. R., Rupp, R. G., & Shoup, V. A. (1999). Characterization of
recombinant human brain-derived neurotrophic factor variants. Archives of biochemistry and
biophysics, 372(2), 248-260.
Sutton, A., & Freundlich, M. (1980). Regulation by cyclic AMP of the ilvB-encoded biosynthetic
acetohydroxy acid synthase in Escherichia coli K12. Molecular and General Genetics MGG,
178(1), 179-183.
Sycheva EV, Yampol’skaya TA, Preobrajenskaya ES, Novikova AE, Matrosov NG, Stoynova NV. 2007.
Overproduction of Noncanonical Amino Acids by Escherichia coli Cells. Microbiology 76(6):
712-718.
Széliová, D., Krahulec, J., Šafránek, M., Lišková, V., & Turňa, J. (2016). Modulation of heterologous
expression from P BAD promoter in Escherichia coli production strains. Journal of
Biotechnology, 236, 1-9.
Tang, Y., & Tirrell, D. A. (2002). Attenuation of the editing activity of the Escherichia coli leucyl-tRNA
synthetase allows incorporation of novel amino acids into proteins in
vivo. Biochemistry, 41(34), 10635-10645.
Tchetina, E., & Newman, E. B. (1995). Identification of Lrp-regulated genes by inverse PCR and
sequencing: regulation of two mal operons of Escherichia coli by leucine-responsive
regulatory protein. Journal of bacteriology, 177(10), 2679-2683.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
157
Tedin, K., & Norel, F. (2001). Comparison of ΔrelA strains of Escherichia coli and Salmonella enterica
serovar Typhimurium suggests a role for ppGpp in attenuation regulation of branched-chain
amino acid biosynthesis. Journal of bacteriology, 183(21), 6184-6196.
Tegel, H., Ottosson, J., & Hober, S. (2011). Enhancing the protein production levels in Escherichia coli
with a strong promoter. FEBS journal, 278(5), 729-739.
Teich, A., Meyer, S., Lin, H. Y., Andersson, L., Enfors, S. O., & Neubauer, P. (1999). Growth rate related
concentration changes of the starvation response regulators σS and ppGpp in glucose‐limited
fed‐batch and continuous cultures of Escherichia coli. Biotechnology progress, 15(1), 123-
129.
Terpe, K. (2006). Overview of bacterial expression systems for heterologous protein production: from
molecular and biochemical fundamentals to commercial systems. Applied microbiology and
biotechnology, 72(2), 211-222.
Thomason, L. C., Sawitzke, J. A., Li, X., Costantino, N., & Court, D. L. (2014). Recombineering: genetic
engineering in bacteria using homologous recombination. Current protocols in molecular
biology, 106(1), 1-39.
Traxler, M. F., Summers, S. M., Nguyen, H. T., Zacharia, V. M., Hightower, G. A., Smith, J. T., &
Conway, T. (2008). The global, ppGpp‐mediated stringent response to amino acid starvation
in Escherichia coli. Molecular microbiology, 68(5), 1128-1148.
Tsai, et al. "Control of mis-incorporation of de novo synthesized norleucine into recombinant
interleukin-2 in E. coli." Biochemical and biophysical research communications 156.2 (1988):
733-739.
Tsui, P., & Freundlich, M. (1990). Integration host factor bends the DNA in the Escherichia coli ilvBN
promoter region. Molecular and General Genetics MGG, 223(2), 349-352.
Umbarger H E. Biosynthesis of the branched chain amino acids. In: Neidhardt F C, et al.,
editors. Escherichia coli and Salmonella: cellular and molecular biology. 2nd ed. Vol. 1.
Washington, D.C.: American Society for Microbiology; 1996. pp. 442–457.
Umbarger, H. E. (1956). Evidence for a negative-feedback mechanism in the biosynthesis of
isoleucine. American Association for the Advancement of Science.
Umbarger, H. E., & Brown, B. (1957). Threonine deamination in escherichia coli ii.: Evidence for two l-
threonine deaminases1. Journal of bacteriology, 73(1), 105.
Usuda, Y., & Kurahashi, O. (2005). Effects of deregulation of methionine biosynthesis on methionine
excretion in Escherichia coli. Applied and environmental microbiology, 71(6), 3228-3234.
Vallejo, L. F., & Rinas, U. (2004). Strategies for the recovery of active proteins through refolding of
bacterial inclusion body proteins. Microbial cell factories, 3(1), 11.
Van Deventer, J. A., Fisk, J. D., & Tirrell, D. A. (2011). Homoisoleucine: a translationally active leucine
surrogate of expanded hydrophobic surface area. ChemBioChem, 12(5), 700-702.
Vartak, N. B., Liu, L., Wang, B. M., & Berg, C. M. (1991). A functional leuABCD operon is required for
leucine synthesis by the tyrosine-repressible transaminase in Escherichia coli K12. Journal of
bacteriology, 173(12), 3864-3871.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
158
References
Veeravalli, K., Laird, M. W., Fedesco, M., Zhang, Y., & Yu, X. C. (2015). Strain engineering to prevent
norleucine incorporation during recombinant protein production in Escherichia
coli. Biotechnology progress, 31(1), 204-211.
Via, P., Badia, J., Baldomá, L., Obradors, N., & Aguilar, J. (1996). Transcriptional regulation of the
Escherichia coli rhaT gene. Microbiology, 142(7), 1833-1840.
Vinogradov, V., Vyazmensky, M., Engel, S., Belenky, I., Kaplun, A., Kryukov, O., ... & Chipman, D. M.
(2006). Acetohydroxyacid synthase isozyme I from Escherichia coli has unique catalytic and
regulatory properties. Biochimica et Biophysica Acta (BBA)-General Subjects, 1760(3), 356-
363
Vitreschak, A. G., Lyubetskaya, E. V., Shirshin, M. A., Gelfand, M. S., & Lyubetsky, V. A. (2006).
Attenuation regulation of amino acid biosynthetic operons in proteobacteria: comparative
genomics analysis. FEMS microbiology letters, 234(2), 357-370.
Wahren, J., Ekberg, K., Johansson, J., Henriksson, M., Pramanik, A., Johansson, B. L., ... & Jrnvall, H.
(2000). Role of C-peptide in human physiology. American Journal of Physiology-
Endocrinology And Metabolism, 278(5), E759-E768.
Walsh, G. (2014). Biopharmaceutical benchmarks 2014. Nature biotechnology, 32(10), 992-1000.
Wang, Q., Wu, J., Friedberg, D., Plakto, J., & Calvo, J. M. (1994). Regulation of the Escherichia coli lrp
gene. Journal of bacteriology, 176(7), 1831-1839.
Wegerer, A., Sun, T., & Altenbuchner, J. (2008). Optimization of an E. coli L-rhamnose-inducible
expression vector: test of various genetic module combinations. BMC biotechnology, 8(1), 2.
Whalen, W. A., & Berg, C. M. (1982). Analysis of an avtA:: Mu d1 (Ap lac) mutant: metabolic role of
transaminase C. Journal of bacteriology, 150(2), 739-746.
Wiegel, J., & Schlegel, H. G. (1977). α-Isopropylmalate synthase from Alcaligenes eutrophus H 16.
Archives of microbiology, 112(3), 239-246
Williams, D. C., Van Frank, R. M., Muth, W. L., & Burnett, J. P. (1982). Cytoplasmic inclusion bodies in
Escherichia coli producing biosynthetic human insulin proteins. Science, 215(4533), 687-689.
Winther-Larsen, H. C., Blatny, J. M., Valand, B., Brautaset, T., & Valla, S. (2000). Pm promoter
expression mutants and their use in broad-host-range RK2 plasmid vectors. Metabolic
engineering, 2(2), 92-103.
Wolfe, A. J. (2005). The acetate switch. Microbiology and molecular biology reviews, 69(1), 12-50.
Wu, G., Yan, Q., Jones, J. A., Tang, Y. J., Fong, S. S., & Koffas, M. A. (2016). Metabolic Burden:
Cornerstones in Synthetic Biology and Metabolic Engineering Applications. Trends in
biotechnology
Xu B, Jahic M, Blomsten G, Enfors SO. 1999a. Glucose overflow metabolism and mixed-acid
fermentation in aerobic large-scale fed-batch processes with Escherichia coli. Applied
Microbiology and Biotechnology 51(5):564-571.
Yoon SH, Han MJ, Jeong H, Lee CH, Xia XX, Lee DH, Shim JH, Lee SY, Oh TK, Kim JF. 2012. Comparative
multi-omics systems analysis of Escherichia coli strains B and K12.Genome. Biol. 13: R37.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
References
159
Young, T. S., & Schultz, P. G. (2010). Beyond the canonical 20 amino acids: expanding the genetic
lexicon. Journal of Biological Chemistry, 285(15), 11039-11044.
Yu, X., Wang, X., & Engel, P. C. (2014). The specificity and kinetic mechanism of branched‐chain
amino acid aminotransferase from Escherichia coli studied with a new improved coupled
assay procedure and the enzyme's potential for biocatalysis. The FEBS journal, 281(1), 391-
400.
Yuan, T., & Vogel, H. J. (1999). Substitution of the methionine residues of calmodulin with the
unnatural amino acid analogs ethionine and norleucine: biochemical and spectroscopic
studies. Protein science, 8(1), 113-121.
Zhu, S., Gong, C., Ren, L., Li, X., Song, D., & Zheng, G. (2013). A simple and effective strategy for
solving the problem of inclusion bodies in recombinant protein technology: His-tag deletions
enhance soluble expression. Applied microbiology and biotechnology, 97(2), 837-845.
Zündorf I, Dingermann T. 2001. Bereitstellung ausreichender Mengen von Humaninsulin. Vom
Rinder-, Schweine-, Pferde-Insulin zum Humaninsulin: Die biotechnische und gentechnische
Insulin-Herstellung. Pharmazie in unserer Zeit 30: 27-32.
Websites
Website 1:
http://www.vectronbiosolutions.com/technologies/
Website 2:
https://nebiocalculator.neb.com/#!/ligation
Website 3:
http://www.clontech.com/US/Support/xxclt_onlineToolsLoad.jsp?citemId=http://bioinfo.clo
ntech.com/infusion/molarRatio.do§ion=16260&xxheight=750
Website 4:
https://www.thermofisher.com/document-connect/document-
connect.html?url=https://assets.thermofisher.com/TFS-
Assets/LSG/manuals/pbaddest49_man.pdf
Website 5:
http://2011.igem.org/Team:DTU-Denmark/Project_improving_araBAD
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
160
Theses
9. Theses
• Genetic regulation of target genes involved in the BCAA biosynthetic pathway affects
biosynthesis and mis-incorporation of ncBCAA into recombinant proteins. Up-regulation of
ilvC, ilvIH or ilvGM and down-regulation of leuA or ilvBN triggers a reduction of ncBCAA
biosynthesis and mis-incorporation. Up-regulation of single operons ilvIH and ilvGM showed
the highest reduction of ncBCAA biosynthesis and mis-incorporation.
• The pyruvate pathway is more likely to redirect more flux to α-ketobutyrate than the
threonine pathway.
• Norvaline, norleucine and by-products of the overflow metabolism accumulate in E. coli
under cultivation conditions subjected to pyruvate pulsing combined with oxygen limitation.
• Metabolic effects triggered in E. coli cultivations subjected to pyruvate pulsing and O2
limitation occur first in the cytosol (intracellular soluble protein fraction), where ncBCAAs are
synthesized. Afterwards, translation machinery mis-incorporates those previously
synthesized ncBCAA present in the cytosol into the nascent recombinant proteins (inclusion
body fraction).
• Norleucine concentration is significantly higher than norvaline in the inclusion body fraction
while β-methylnorleucine could not be detected. This suggests that the amount of a certain
cBCAA in the amino acid sequence of a recombinant protein is not the main factor
determining the mis-incorporation probability of the respective ncBCAA analog in such
recombinant protein.
• Despite being norvaline and norleucine present in similar concentration levels intracellularly,
the probability of misaminoacylation by met-tRNA seems to be higher than by leu-tRNA.
• Norvaline and norleucine concentrations were higher at 15 L reactor than at mini-reactor
scale. Although not comparable to industrial scale reactors, it might be that mixing in the 15L
reactor is inefficient, thus triggering formation of O2 and glucose gradients, which, in turn,
lead to ncBCAA formation, even under standard cultivation conditions.
• IPTG-mediated induction of recombinant protein expression is the main factor triggering
ncBCAA formation under standard cultivation conditions.
• Neither genetic regulation of target genes involved in the BCAA biosynthetic pathway nor
cultivation conditions subjected to pyruvate pulsing and oxygen limitation caused a
significant alteration of β-methylnorleucine levels in E. coli. This may suggest that β-
methylnorleucine is synthesized by an alternative unknown metabolic pathway.
• For each target gene, different induction strength, i.e. L-arabinose concentration, is
necessary in order to trigger genetic expression levels enough to recover cell growth levels of
the wild type strain. This might be explained due to the different degree of essentiality of the
target genes for the cell. Lack of one of the three ilvGM, ilvIH and ilvBN gene products does
not dramatically affect growth behavior since, as isoenzymes, its function can be also carried
out by the other two available gene products.
• Induction efficiency of lac-based promoters present in expression plasmids is strongly
dependent on the levels of LacI repressor endogenously expressed in the E. coli strain
selected as host for expression.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
161
10. Appendix
Table S1. Reagents used in this study.
Category
Name
Cataloge number
Manufacturer
Antibiotics
Ampicillin sodium salt
A9518
Sigma
Chloramphenicol
C0378
Sigma
Kanamycin disulfate salt
K1876
Sigma
Amino acids
L-norleucine
N6877
Sigma
L-norvaline
N7627
Sigma
ß-methylnorleucine
4036488
Bachem
L-2-aminobutyric acid
A1879
Sigma
Enzymes
Amylase Reagent A 3000 U/L
-
Biosilta
Lysonase Bioprocessing Reagent
71230
Merck
Dreamtaq polymerase
EP0702
Thermoscientific
Phusion DNA Polymerase
M0530
NEB
Pfu polymerase
EP0571
Thermoscientific
XhoI
R0146
NEB
MssI
R0560
NEB
NheI-HF
R3131
NEB
NotI-HF
R3189
NEB
PstI-HF
R3140
NEB
HindIII-HF
R3104
NEB
DpnI
R0176
NEB
EcoRI-HF
R3101
NEB
PaeI-HF
R3182
NEB
PciI
R0655
NEB
Quick ligase
M2200
NEB
Exo-Star Reagent
GEUS78210
Sigma
Protein and DNA
Quick-Load® 2-Log DNA Ladder
N0469S
NEB
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
162
Appendix
markers
SeeBlue® Plus2 Pre-stained
Protein Standard
LC5925
Thermo Fisher Scientific
Protein and nucleic
acid stains
Ethidium bromide, 10 mg/mL
15585-011
Invitrogen
GelRed Nucleic Acid Gel Stain
10,000X stock solution
41003
Biotium
Instant Blue
ISB1L
Sigma-Aldrich
SDS-PAGE reagents
NuPAGE LDS Sample Buffer (4X)
NP0007
Thermo Fisher Scientific
NuPAGE MES SDS Running Buffer
(20X)
NP0002
Thermo Fisher Scientific
NuPAGE Sample Reducing Agent
(10X)
NP0009
Thermo Fisher Scientific
NuPAGE Antioxidant
NP0005
Thermo Fisher Scientific
Bolt LDS Sample Buffer (4X)
B0008
Thermo Fisher Scientific
Bolt MES SDS Running Buffer
(20X)
B0002
Thermo Fisher Scientific
Bolt Sample Reducing Agent
(10X)
B0009
Thermo Fisher Scientific
Bolt Antioxidant
B0005
Thermo Fisher Scientific
NuPAGE 12% Bis-Tris Plus gels
NP0342BOX
Thermo Fisher Scientific
Bolt 12% Bis-Tris Plus gels
NW00120BOX
Thermo Fisher Scientific
Components for DNA
electrophoresis
6X Gel Loading Dye
B7024S
NEB
E-Gel® 1.2 % Agarose (EtBr
stained)
G501801
Thermoscientific
Agarose
840004
Biozym
PCR reagents
Thermopol-buffer 10X
B9004
NEB
10X DreamTaq buffer
B65
Thermoscientific
5X Phusion GC Buffer
M0530
NEB
5X Phusion HF Buffer
M0530
NEB
100% DMSO
B0515A
NEB
MgCl2
M0530
NEB
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
163
dNTPs 10mM
N0447
NEB
Buffers
10 mM Tris/1 mM EDTA (TE
buffer)
93283
Sigma
10X CutSmart Buffer
B7204
NEB
Quick ligase reaction buffer 2X
M2200
NEB
10X TAE Buffer
15558-042
Gibco
Kits
Quick ligation kit
M2200
NEB
In-Fusion® HD Cloning Kit
638909
Takara
QIAGEN Plasmid Plus Midi Kit
ID12943
Qiagen
QIAprep® Miniprep kit
27106
NEB
QIAquick Gel Extraction Kit
28706
Qiagen
QIAquick PCR Purification Kit
28106
Qiagen
EZ:faastTM for free (physiological)
amino acid analysis by GC-FID kit
KG0-7165
Phenomenex
BugBuster Protein Extraction
Reagent
70584-4
Merck
Reagents for media
preparation
D(+)-Glucose monohydrate
108342
Merck
MgSO4*7H2O
M2773
Sigma
Invitrogen UltraPure™ Distilled
Water
10977-035
Thermo Scientific
LB-Agar (Lennox)
X965.1
Roth
LB-Medium (Lennox)
X964.1
Roth
SOC medium Outgrowth
Medium
B9020S
NEB
FeCl3*6H2O
236489
Sigma
CaCl2*2H2O
C3306
Sigma
MnCl2*4H2O
M8054
Sigma
ZnSO4*7H2O
Z0251
Sigma
CoCl2*6H2O
C8661
Sigma
CuCl2*2H2O
459097
Sigma
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
164
Appendix
NiCl2*6H2O
13613
Sigma
Na2MoO4*5H2O
331058
Sigma
H3BO3
31146
Sigma
K2HPO4
1551128
Sigma
KH2PO4
NIST200B
Sigma
Na citrate*2H2O
W302600
Sigma
(NH4)2SO4
A4418
Sigma
NaCl
S9625
Sigma
Thiamine hydrochloride
T4625
Sigma
BactoTM Casamino acids
223050
BD
Na2HPO4*2H2O
1.06580.0500
Merck
NH4Cl
1.011143.0050
Merck
Na2SO4
31481
Sigma
NaH2PO4
S0751
Sigma
(NH4)2-H-citrate
09833
Sigma
MnSO4*H2O
1.05941.0250
Merck
CuSO4*5H2O
31293
Sigma
Reagents for media
supplementation
L-arabinose
A3256
Sigma
EnPump 200 substrate
-
Enpresso
Sodium pyruvate
P2256
Sigma
Desmophen
-
Covestro
Antifoam 204
A6426
Sigma
25 % NH4OH solution
06010.4010
Bernd Kraft
m-toluate
T36609
Sigma
IPTG
I6758
Sigma
Others
Hydrochlorhydric acid 5M
10605882
Fischer Scientific
Isopropanol
184130010
Acros Organics
Glycerol
535036
Hedinger
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
165
Betaine
B24397
Alfa Aesar
Bardac
-
Lonza
Table S2. Materials used in this study.
Material
Catalog number
Manufacturer
Eppendorf tubes 1.5 mL
0030120.086
Eppendorf
Eppendorf tubes 2 mL
0030123.344
Eppendorf
Eppendorf tubes 5 mL
0030119.401
Eppendorf
PCR tubes 0.2 mL
0030124.332
Eppendorf
PCR tube Stripes 0.2 mL (used for
sequencing)
0030124.359
Eppendorf
Nalgene Cryogenic vials (2 mL)
5000-0020
Thermo Scientific
Gene Pulser®/MicroPulser™
Electroporation Cuvettes, 0.2 cm gap
1652086
Biorad
15 mL round-bottom tube
0030122.151
Eppendorf
50 mL Falcon tubes
0030122.178
Eppendorf
24-well deep well plates (Pall mini-
reactor)
MRT-PRC-21
Pall
96-well microplate
04-083-0150
Nerbe plus
Syringe filter 0.45 µm
10462100
Healthcare Life Sciences
Syringe filter 0.22 µm
10462200
GE Healthcare Life Sciences
Syringes
Omnifix Syringes
B.Braun
Canules
Sterican
B. Braun
Pipette tips
epT.I.P.S.
Eppendorf
Pipettes kit
Research plus
Eppendorf
Pipette filler
Accu-jet pro
Brand
Serological pipettes
Nunc serological pipette
Thermo Scientific
Filtration unit (diff. volumes)
Express Plus 0.22 µm
(SCGPU11RE)
Millipore
Steriflip®
SCGP00525
Millipore
Cuvette PMMA, semi-micro (1.5 mL)
759115
Brand
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
166
Appendix
Inoculation loop 1 µL
731165
Fischer Scientific
Inoculation loop 10 µL
731175
Fischer Scientific
Inoculation needle
731185
Fischer Scientific
L-shaped spreader
174CS05
Copan
Table S3. Equipment used in this study.
Equipment
Model
Manufacturer
Thermocycler
SimpliAmp Thermal Cycler
Thermo Scientific
Electroporator
Gene Pulser Xcell
Biorad
Spectrophotometer
NanoDrop One
Thermo Scientific
Photometer
Ultraspec 2100 pro
Amersham Bioscience
Gel documentation
Gel Doc EZ Imager
Biorad
Sonotrode
VialTweeter UP200St
Hielscher
GC-FID instrument
7890A
Agilent Technologies
GC Autoinjector
CombiPAL Injector CTC Analytics
G6501-CTC
Agilent Technologies
Hydrogen generator
500
Schmidlin
GC column
ZB-AAA Zebron Amino Acid GC
Column 10mx0.25mm CG0-7169
Phenomenex
Mini-reactor system
Pall Micro24
Microreactor Technologies Inc.
15L reactor
Type 880142.8, Nr. 209
Braun Melsungen
DCU reactor
BiostatR ED
B. Braun Biotech International
BioPAT Trace
biopattrace
Sartorius
Analytical balance
ME204T/00
Mettler Toledo
Balance
Laboratory LC6200S
Sartorius
Centrifuge
75005510/01
Biofuge Fresco
Centrifuge
Heraeus Multifuge X1R
Thermo Scientific
Centrifuge
Heraeus Pico 21
Thermo Scientific
Vortex
444-1372
VWR International
DNA electrophoresis units
E-Gel iBase™ Power System
Thermo Scientific
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
167
Thermomixer
88880028
Thermo Scientific
Vertical shaker
Trayster D
IKA
Incubator
IPP 55 PLUS
Memmert
Incubator
HERATHERM Incubator
Thermo Scientific
pH Meter
Multi 9310 IDS
WTW
Protein electrophoresis unit
XCell SureLoc Mini-Cell
Thermo Scientific
Horizontal gel electrophoresis
chamber
460.000 (midi large)
-
Horizontal gel electrophoresis
chamber
MSMINIDUO
Biozym
Water bath
WNB22
Memmert
UV Transilluminator
UST-20M-8R (Transilluminator
BioView)
Biostep
Laminar flow bench
50067140
Heraeus
Magnetic stirrer
RSM-10HS
Phoenix Instrument
Microwaves
MM817ALR
Siemens
Incubator shaker
Kuhner SHAKER X / Climo-shaker
ISF1-X
Kuhner
Incubator shaker
886342/3 (Certomat R)
B.
Table S4. Software used in this study.
Software
Provider
Application
SnapGene
GSL Biotech LLC
Primer design, generation of
plasmid maps, aligment of
sequences, in silico molecular
verification.
NEBio® Claculator
NEB
Calculation of ligation parameters.
Tm Calculator
NEB
Determination of Ta of primers for
PCR.
In-Fusion® Molar Ratio Calculator
Takara
Calculation optimal amounts of
vector and insert for the In-Fusion
Cloning reaction.
Multiple Primer Analyzer
Thermo Scientifc
Analysis and comparison of
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
168
Appendix
multiple primer sequences.
Lucullus PIMS v 3.5.2
Securecell
Process monitoring during
fermentations.
Image Lab
Biorad
Acquisition and anlysis of DNA and
protein gel images.
Agilent Chemstation, Rev. B.03.02
[341]
Agilent Technologies
Programming of GC runs and
chromatogram analysis.
Leica Application Suite V4
Leica Microsystems
Acquisition, storage, annotation
and display of microscope images.
Table S5. Strains used in this study and its features.
Strain
Genotype
Source
Comment
E. coli K-12
NEB5α
fhuA2, (argF-lacZ)U169, phoA,
glnV44, 80(lacZ)M15, gyrA96, recA1,
relA1, endA1, thi-1, hsdR17
NEB (Cat. Nr.:
C2987H)
Used as recipient strain for
plasmid storage.
E. coli K-12
W3110M
F-, λ-, IN(rrnD-rrnE)1, rph-1, ilvG-1,
lacIq
Internal
E. coli wild type strain. Used as
recipient strain.
E. coli K-12
BW25113
F-, Δ(araD-
araB)567, ΔlacZ4787(::rrnB-3), λ-
, rph-1, ilvG-1, Δ(rhaD-
rhaB)568, hsdR514, lacI+
CGSC#7636
E. coli wild type strain mainly
used in this study. Used as
recipient strain.
E. coli K-12
BW25113
pKD46
See E. coli K-12 BW25113
CGSC#7739
Strain containing plasmid
pKD46, necessary for the
recombineering process used in
this study.
E. coli K-12
BW25141
pKD3
F-, Δ(araD-
araB)567, ΔlacZ4787(::rrnB-
3), Δ(phoB-phoR)580, λ-
, galU95, ΔuidA3::pir+, recA1, endA9(
del-ins)::FRT, rph-1, Δ(rhaD-
rhaB)568, hsdR51
CGSC#7631
Strain containing plasmid pKD3,
necessary for the
recombineering process used in
this study.
E. coli K-12
BW25141
pKD4
See E. coli K-12 BW25141 pKD3
CGSC#7632
Strain containing plasmid pKD4,
necessary for the
recombineering process used in
this study.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
169
E. coli K-12
BT340 pCP20
F-, Δ(argF-lac)169,
φ80dlacZ58(M15), glnX44(AS), λ-,
rfbC1, gyrA96(NalR), recA1, endA1,
spoT1, thiE1, hsdR17
CGSC#7629
Strain containing plasmid
pCP20, necessary for the
recombineering process used in
this study.
E. coli K-12
BW25113
leuA:kanR
pKD46*
See E. coli K-12 BW25113
CGSC#JW0073-1
leuA KO-mutant containing
plasmid pKD46. Kanamycin
resistance cassette is
integrated in leuA genetic
region.
E. coli K-12
BW25113
ilvA:kanR
pKD46*
See E. coli K-12 BW25113
CGSC#JW3745-2
ilvA KO-mutant containing
plasmid pKD46. Kanamycin
resistance cassette is
integrated in ilvA genetic
region.
E. coli K-12
BW25113
ilvC:kanR
pKD46*
See E. coli K-12 BW25113
CGSC#JW3747-2
ilvC KO-mutant containing
plasmid pKD46. Kanamycin
resistance cassette is
integrated in ilvC genetic
region.
E. coli K-12
BW25113
thrA:kanR
pKD46*
See E. coli K-12 BW25113
CGSC#JW0001-1
thrA KO-mutant containing
plasmid pKD46. Kanamycin
resistance cassette is
integrated in thrA genetic
region.
E. coli K-12
BW25113
ΔilvIH
See E. coli K-12 BW25113
This study
ilvIH KO-mutant. Used as
recipient strain.
E. coli K-12
BW25113
ΔilvBN
See E. coli K-12 BW25113
Sarah Charaf
ilvBN KO-mutant. Used as
recipient strain.
*: These strains were further processed as follows in order to obtain the recipient strain: plasmid pKD46 was
curated from the acquired E. coli K-12 BW25113 single knock-out mutants and the respective electro-
competent cells were transformed with plasmid pCP20 in order to trigger removal of the kanamycin cassette
from the genome by FRT-specific recombination. Plasmid pCP20 was then curated giving the corresponding
leuA, thrA, ilvA and ilvC KO-recipient strains.
Table S6. Plasmids used in this study and its features.
Plasmid
Size (bp)
Antibiotic
resistance
Source
Expressed
protein
Features
pKD3
2804
amp, cm
CGSC#7631
-
R6K ori, 3 priming sites for
PCR, cmR flanked by FRT
sites
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
170
Appendix
pKD4
3267
amp, kan
CGSC#7632
-
R6K ori, 3 priming sites for
PCR, kanR flanked by FRT
sites
pKD46
6329
amp
CGSC#7739
lambda (λ) Red
recombination
system (Proteins
Exo, Bet
und Gam)
araBAD promoter, pSC101
ori (Rep101)
pCP20
9497
amp, cm
CGSC#7629
Flippase (for FRT-
specific
recombination)
pSC101 ori (Rep101)
pSW3
3609
amp
Internal
Mini-proinsulin
Ptac promoter, pMB1
(derivative) ori
pSW3_lacI+
3982
amp
This study
Mini-proinsulin
and LacI
repressor
Ptac promoter, pMB1
(derivative) ori, LacI (lacI+
promoter)
pSW3_lacIq
3982
amp
This study
Mini-proinsulin
and LacI
repressor
Ptac promoter, pMB1
(derivative) ori, LacI (lacIq
promoter)
pET-coco1
12474
cm
Novagen
(Cat. Nr.:
71129)
-
T7 promoter, oriV (TrfA,
araBAD promoter), oriS
(RepE, SopA, SopB, SopC)
pACG_araBAD
7237
cm
This study
-
araBAD promoter, oriS
(RepE, SopA, SopB, SopC)
pACG_araBAD_
thrA
9680
cm
This study
Bifunctional
aspartokinase/ho
moserine
dehydrogenase 1
See pACG_araBAD
pACG_araBAD_i
lvA
8762
cm
This study
L-threonine
dehydratase
biosynthetic
See pACG_araBAD
pACG_araBAD_l
euA
8789
cm
This study
2-isopropylmalate
synthase
See pACG_araBAD
pACG_araBAD_i
lvIH
9436
cm
This study
Acetolactate
synthase isozyme
3
See pACG_araBAD
pACG_araBAD_i
lvBN
9200
cm
This study
Acetolactate
synthase isozyme
1
See pACG_araBAD
pACG_araBAD_i
lvGM
9121
cm
This study
Acetolactate
synthase isozyme
2
See pACG_araBAD
pACG_araBAD_i
lvC
8693
cm
This study
Ketol-acid
reductoisomerase
(NADP(+))
See pACG_araBAD
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
171
pACG_XylSPm
7948
cm
This study
-
XylSPm promoter, oriS
(RepE, SopA, SopB, SopC)
pACG_XylSPm
_thrA
10391
cm
This study
Bifunctional
aspartokinase/ho
moserine
dehydrogenase 1
See pACG_XylSPm
pACG_ XylSPm
_ilvA
9473
cm
This study
L-threonine
dehydratase
biosynthetic
See pACG_XylSPm
pACG_ XylSPm
_leuA
9500
cm
This study
2-isopropylmalate
synthase
See pACG_XylSPm
pACG_ XylSPm
_ilvC
9404
cm
This study
Ketol-acid
reductoisomerase
(NADP(+))
See pACG_XylSPm
pACG_ XylSPm
_ilvBN
9911
cm
This study
Acetolactate
synthase isozyme
1
See pACG_XylSPm
16ABZ5NP_193
4177_araBAD_
MCS1
3876
kan
GeneArt
-
ColE1 ori
16ADFSUP_203
4902_lacIq_pro
motor
4214
kan
GeneArt
-
ColE1 ori
17ABNF2P_211
5088_sacBopt
3709
kan
GeneArt
Levansucrase
(codon-optimized
for E. coli)
ColE1 ori
16ADCJKP_202
8165_XylSPm
4461
kan
GeneArt
-
ColE1 ori
Table S7. Primers used in this study and its application.
Primer name
Sequence 5` → 3`
Primer application
pSW3_F1_seq
AACCTTTCGCGGTATGGCATG
1. Sequencing of plasmid pSW3.
2. Primers pSW3_R3_seq and
pSW3_R6_seq were additionally
used for plasmid verification by
colony PCR.
pSW3_R1_seq
GGGCGCTATCATGCCATAC
pSW3_R2_seq
TTGATGGGTGTCTGGTCAGAGAC
pSW3_F2_seq
ATGAAGACGGTACGCGACTG
pSW3_F3_seq
GCGTGGTATCGTTGAACAATG
pSW3_R3_seq
CAACATTGTTCAACGATACCAC
pSW3_F4_seq
GACTCAACTGCAACTGGAACAC
pSW3_R4_seq
CGAAACGGCCCTCAATCATAC
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
172
Appendix
pSW3_F5_seq
CCGCTTTATCAGAAGCCAGAC
pSW3_R5_seq
GCTACGTGACTGGGTCAT
pSW3_F6_seq
CCAGCAAAAGGCCAGGAAC
pSW3_R6_seq
GACCTACACCGAACTGAGATAC
pSW3_F7_seq
TCTGCTGAAGCCAGTTACCTTC
pSW3_R7_seq
GAGTCAGGCAACTATGGATGAAC
pSW3_F8_seq
GCAGGCATCGTGGTGTCAC
Mut_lacI+_F
P-CATCGAATGGCGCAAAACCTTTCGC
Mutagenesis of plasmid
pSW3_lacIq to generate
pSW3_lacI+.
Mut_lacI+_R
P-GTGTCGCATGCCGCTTCG
araBAD_InFusion_F
CTCCCGGCATCCGCTTACAGACAAG
PCR from plasmid
16ABZ5NP_1934177 to generate
Fragment 1, which participates in
the InFusion cloning process to
generate plasmid pACG_araBAD.
araBAD_InFusion_R
CAAAAAACCCCTCAAGACCCGTTTAGAG
cmR_InFusion_F
TTGAGGGGTTTTTTGATTTAAATAGCGCTAACCGT
TTTTATCAGGCTCTG
PCR from plasmid pCP20 to
generate Fragment 2, which
participates in the InFusion
cloning process to generate
plasmid pACG_araBAD.
cmR_InFusion_R
CACCTAGGCTCGAGTGATCGGCACGTAAGAGGTT
CCAAC
ori2_InFusion_F
ACTCGAGCCTAGGTGAGCGAGGAAGCACCAGGG
AACAG
PCR from plasmid pETcoco1 to
generate Fragment 3, which
participates in the InFusion
cloning process to generate
plasmid pACG_araBAD.
ori2_InFusion_R
AGCGGATGCCGGGAGGTTTAAACAAGCAGGACA
CAGCAGCAATCCACAG
InFusion1_seq_F1
GAGCTGCATGTGTCAGAGGTTTTC
PCR-verification and sequencing
of plasmid pACG_araBAD and
pACG_XylSPm.
InFusion1_seq_F2
GCGACAAGCAAACATGCTGTG
InFusion1_seq_F3
TCGGCAAACAAATTCTCGTCCCTG
InFusion1_seq_F4
AAGATTAGCGGATCCTACCTGAC
InFusion1_seq_F5
AAGCGGCTATTTAACGACCCTG
InFusion1_seq_F6
ATCGTCGTGGTATTCACTCCAG
InFusion1_seq_F7
ACCTCTTACGTGCCGATCACTC
InFusion1_seq_F8
AATAGCCCGCGAATCGTCCAG
InFusion1_seq_F9
AGTATCGTAAGCCGGATGGCTC
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
173
InFusion1_seq_F10
TCTGAATTGGCTATCCGCGTG
InFusion1_seq_F11
ATAACCCGGCGCTGGAGAATAG
InFusion1_seq_F12
ATGCAATAAAGCCCACTTGCTG
InFusion1_seq_F13
GGTAAAGGTCAGATCCGGATGAG
InFusion1_seq_F14
TGGTCAACAGACACCGGCGTTC
InFusion1_seq_F15
AACGTCATCTGCATCAAGAACTAG
InFusion1_seq_F16
CTGGGCCCACTGTTCCACT
InFusion1_seq_F17
GTCGATCAGACTATCAGCGTGAG
InFusion1_seq_R1
GTTCCGTGCCGGTTGTGAAG
InFusion1_seq_R2
AGCGACCTGTTTGGGCAAATC
InFusion1_seq_R3
TTCTTCTCTGAATGGCGGGAG
InFusion1_seq_R4
ACCCCTCAAGACCCGTTTAGAG
InFusion1_seq_R5
CAGGTTCATCATGCCGTCTGTG
InFusion1_seq_R6
CCTATAACCAGACCGTTCAGCTG
InFusion1_seq_R7
ACACCTTCTCTAGAACCAGCATG
InFusion1_seq_R8
ATCGCCGGCATCCTCTTCAG
InFusion1_seq_R9
TGTGACGAACCACCCTCAAATC
InFusion1_seq_R10
GCAACAACAAAATCGCAAAGTCATC
InFusion1_seq_R11
CGATCACCGGTGGAAATACGTC
InFusion1_seq_R12
AGTCAAACAACTCAGCAGGCGTG
InFusion1_seq_R13
AGCTGGTGTCGATAACGAAGTATC
InFusion1_seq_R14
AATTCATTCTGCAATCGGCTTG
InFusion1_seq_R15
ATTCAGGCCAGTTATGCTTTCTG
InFusion1_seq_R16
AGACCGACAACACGAGTGGGATC
thrA_NheI_F
AGCAGCTAGCATGCGAGTGTTGAAGTTCGGCGGT
AC
PCR of native gene thrA from E.
coli K-12 BW25113 genomic DNA
for subsequent NheI- and NotI-
mediated cloning into plasmid
variants pACG_araBAD and
pACG_XylSPm.
thrA_NotI_R
ATTAGCGGCCGCCGACTCCTAACTTCCATGAGAG
GGTACGTAGCAGATC
ilvA_NheI_F
AGCAGCTAGCATGGCTGACTCGCAACCCCTGTCC
PCR of native gene ilvA from E.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
174
Appendix
GGTGCTC
coli K-12 BW25113 genomic DNA
for subsequent NheI- and NotI-
mediated cloning into plasmid
variants pACG_araBAD and
pACG_XylSPm.
ilvA_NotI_R
ATTAGCGGCCGCCACCCGCCAAAAAGAACCTGAA
CGCCGGGTTATTG
leuA_NotI_F
ATTAGCGGCCGCCCACGGTTTCCTTGTTGTTTTCG
TTG
PCR of native gene leuA from E.
coli K-12 BW25113 genomic DNA
for subsequent NheI- and NotI-
mediated cloning into plasmid
variants pACG_araBAD and
pACG_XylSPm.
leuA_NheI_R
AGCAGCTAGCATGAGCCAGCAAGTCATTATTTTC
GATACCACATTG
ilvGM_NotI_F
AGCAGCTAGCATGAATGGCGCACAGTGGGTGGT
ACATGCGTTG
PCR of native operon ilvGM from
E. coli K-12 BW25113 ilvG+
genomic DNA for subsequent
NheI- and NotI-mediated cloning
into plasmid variants
pACG_araBAD and
pACG_XylSPm.
ilvGM_NheI_R
ATTAGCGGCCGCCGGCGCGGATTTGTTGTGATGT
GGTTGTGCTCTG
ilvIH_NheI_F
AGCAGCTAGCATGGAGATGTTGTCTGGAGCCGAG
ATGGTCGTC
PCR of native operon ilvIH from
E. coli K-12 BW25113 genomic
DNA for subsequent NheI- and
NotI-mediated cloning into
plasmid variants pACG_araBAD
and pACG_XylSPm.
ilvIH_NotI_R
ATTAGCGGCCGCCACGCATTATTTTATCGCCGCGC
GAAAGTCCGAC
ilvC_NheI_F
AGCAGCTAGCATGGCTAACTACTTCAATACACTGA
ATCTGCGCCAG
PCR of native gene ilvC from E.
coli K-12 BW25113 genomic DNA
for subsequent NheI- and NotI-
mediated cloning into plasmid
variants pACG_araBAD and
pACG_XylSPm.
ilvC_NotI_R
ATTAGCGGCCGCCACCCGCAACAGCAATACGTTT
CATATCTG
ilvBN_NheI_F
ATTAGCGGCCGCCCTGAAAAAACACCGCGATCTT
GTTAAACATC
PCR of native operon ilvBN from
E. coli K-12 BW25113 genomic
DNA for subsequent NheI- and
NotI-mediated cloning into
plasmid variants pACG_araBAD
and pACG_XylSPm.
ilvBN_NotI_R
AGCAGCTAGCATGGCAAGTTCGGGCACAACATCG
AC
leuA_F1_seq
GGTGATCGAACGCGCTATCTATATG
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_leuA and
pACG_XylSPm_leuA
leuA_F2_seq
CAGCCAGTTAGTTAGCCAGATTTG
leuA_F3_seq
GTCCGGTCGATGCCGTCTATC
leuA_R1_seq
CGTCAACACCCATACGCTCAAG
leuA_R2_seq
GGTATGTACGGAGATAATGGCTTTG
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
175
leuA_R3_seq
CAGGAAAGCATCGTACAAATTGTC
ilvC_F1_seq
CGTACAGCCACTGATGAAAGAC
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_ilvC and
pACG_XylSPm_ilvC
ilvC _F2_seq
GATGATGGACCGTCTCTCTAACC
ilvC _F3_seq
GCTGAAACCGTTTATGGCAGAG
ilvC _R1_seq
CATGGTTTCGAACGCCAGTTC
ilvC _R2_seq
TTCACTTCCGCAACGAAGGAC
ilvC _R3_seq
TTCAGACCCTGTGCGCCACAG
ilvA_F1_seq
ATCGTTATGCCAACCGCCAC
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_ilvA and
pACG_XylSPm_ilvA
ilvA _F2_seq
GTCGATAGCGATGCGATCTGTG
ilvA _F3_seq
GCTACAGCGTGGTTGATC
ilvA _R1_seq
GCACACCGACAAAGATGCAG
ilvA _R2_seq
CATCCAGCGCTGCTTTCAG
ilvA _R3_seq
GCTTTCTGTTCTTCCGTCAGG
ilvIH_F1_seq
CTTTCAGGAGTGCGACATGG
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_ilvIH and
pACG_XylSPm_ilvIH
ilvIH _F2_seq
GAATGCACGGTACCTACG
ilvIH _F3_seq
CATTCGACAAACCGCGTC
ilvIH _F4_seq
GTCTACCCGATGCAGATTCG
ilvIH _F5_seq
ATGTCACACCCTCGCTTTATAC
ilvIH _R1_seq
CACGGTCTGGATGGTCATAC
ilvIH _R2_seq
CGTATTGCAACGCGGTAGAC
ilvIH _R3_seq
GGCGGATTCTTGCGACAAG
ilvIH _R4_seq
GTGGGATTGTAAGAACGCATACTG
ilvIH _R5_seq
GCCTCCGGGATAACCGAATAC
thrA_F1_seq
CAGTGCCCGGATAGCATCAAC
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_thrA and
pACG_XylSPm_thrA
thrA _F2_seq
CCATCGCCCAGTTCCAGATC
thrA _F3_seq
GGATCTTCTGAACGCTCAATCTC
thrA _F4_seq
CTCGTCGATGGATTACTACCATC
thrA _F5_seq
ATTGATGAAGATGGCGTCTGC
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
176
Appendix
thrA _R1_seq
ATCAATAGTTTACGCGCCACATC
thrA _R2_seq
CTCTTTGGCTTGCGCCAGTTC
thrA _R3_seq
CGTTCAGCTCGCACACAGTC
thrA _R4_seq
AGCAGAGTAGTCGGAACCGTTG
thrA _R5_seq
TAAAGCATCCTGGCCGCTAATG
ilvGM_F1_seq
CACCGTTTATCGGCACTGAC
PCR-verification and sequencing
of plasmid variants
pACG_araBAD_ilvGM and
pACG_XylSPm_ilvGM
ilvGM_F2_seq
GCACCAAAGCGGCAAACTTC
ilvGM_F3_seq
CTCCAGCGGTTTAGGTACCATG
ilvGM_F4_seq
ATGTATCGGCTCGCTTCAATC
ilvGM_R1_seq
GAAGGCGCTGGCTAACATGAG
ilvGM_R2_seq
TCACCGGGATGGTCGTAAC
ilvGM_R3_seq
ACCGCCAACGTACAGCATC
ilvGM_R4_seq
GTAGCACGAGCATAACCGATAG
ilvBN_R2_seq
GCCATAGCGGGCTGTGTTTC
PCR-verification and sequencing
of plasmid pACG_XylSPm_ilvBN
ilvBN_F1_seq
TGGACACCTACGGCATCTCTATC
KO3_F
TTTACACATTTTTTCCGTCAAACAGTGAGGCAGGC
CGTGTAGGCTGGAGCTGCTTC
PCR from plasmids pKD3 or pKD4
to generate the ilvIH knock-out
cassette.
KO3_R
ACATGTTGGGCTGTAAATTGCGCATTGAGATCATT
CATGGGAATTAGCCATGGTCC
ilvIH_F
TACCGCTGCCGTTAAAACAAAACAG
PCR-verification and sequencing
of the target genomic ilvIH
region.
ilvIH_R
CCAGTTTCACAATTGCCCCTTGC
ilvIH_F2
ATGGCAATGATGAGCGTACTTAGCGTGG
ilvA_F
TTACAGGTAAGCGATGCCGAACTGG
PCR-verification of the target
genomic ilvA region
ilvA_R
GGTGCGTAATCAGGTGTCGGTAGAG
thrA_F
GTTCTTACTGGTGTGCCGATGGTGG
PCR-verification of the target
genomic thrA region
thrA_R
CCAGCAAACGAGTGTCATTAAGCGG
leuA_F
GGGCGCAGGTTGCTGAATAATTTG
PCR-verification of the target
genomic leuA region
leuA_R
TTTACGGATGCAGAACTCACGCTG
ilvC_F
GGATCATGCTAAAATCGCCGCCTG
PCR-verification of the target
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
177
ilvC_R
CGCGTAAATCCACAGGGACAGC
genomic ilvC region
Figure S1. Plasmid maps of pKD3 (A), pKD4 (B), pKD46 (C) and pCP20 (D). The use of these plasmids allows
homologous recombination in E. coli. Plasmid maps were generated by Snapgene®.
A B
C D
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
178
Appendix
Figure S2. Plasmid maps of pSW3 (A), pSW3_lacI+ (B) and pSW3_lacIq (C). Plasmid maps were generated by
Snapgene®.
A
B C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
179
Figure S3. Plasmid maps of 16ADFSUP_2034902_lacIq_promotor (A), 16ABZ5NP_1934177 (B),
16ADCJKP_2028165_XylSPm (C). These plasmids were provided by the GeneArt® Gene Synthesis service of
Termofisher scientific. Plasmid maps were generated by Snapgene®.
A
B C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
180
Appendix
Figure S4. Plasmid map of pGEX-4T-1 (A), pETcoco-1 (B), pBAD_DEST49 (C) and pJB658 (D). Plasmid maps were
generated by Snapgene®.
A B
C D
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
181
A B
C D
E F
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
182
Appendix
Figure S5. Plasmid map of the different pACG_araBAD plasmid variants generated in this study: pACG_araBAD
(A), pACG_araBAD_ilvA (B), pACG_araBAD_thrA (C), pACG_araBAD_leuA (D), pACG_araBAD_ilvBN (E),
pACG_araBAD_ilvGM (F), pACG_araBAD_ilvIH (G) and pACG_araBAD_ilvC (H). Plasmid maps were generated by
Snapgene®.
G H
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
183
A B
C D
E F
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
184
Appendix
Figure S6. Plasmid map of the different pACG_XylSPm plasmid variants generated in this study: pACG_XylSPm
(A), pACG_XylSPm_thrA (B), pACG_XylSPm_ilvA (C), pACG_XylSPm_leuA (D), pACG_XylSPm_ilvBN (E),
pACG_XylSPm_ilvIH (F) and pACG_XylSPm_ilvC (G). Plasmid maps were generated by Snapgene®.
G
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
185
Figure S7. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pSW3_lacIq clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for E. coli
K-12 NEB5α pSW3_lacIq is 2010 bp (indicated in red). 2-20: clones 2-20 potential E. coli K-12 NEB5α pSW3_lacIq;
M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S8. Gel electrophoresis of PCR products obtained by mutagenesis PCR of plasmid pSW3_lacIq by using
primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size is 3982 bp (indicated by a red
arrow). Four different buffer combinations were tested for PCR. 1: buffer HF; 2: buffer HF plus DMSO; 3: buffer
GC; 4: buffer GC plus DMSO; 5: negative control; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder
(0.1-10.0 kb).
M 11 12 13 14 15 16 17 18 19 20 M
M 2 3 4 5 6 7 8 9 10 M
M 1 2 3 4 5 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
186
Appendix
Figure S9. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for E. coli
K-12 NEB5α pSW3_lacI+ is 2010 bp (indicated by a red arrow). 1-9: clones 1-9 potential E. coli K-12 NEB5α
pSW3_lacI+; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S10. Gel electrophoresis of EcoRI-restriction products from plasmid preparations of potential E. coli K-12
BW25113 pSW3 clones for verification of plasmid pSW3 presence. Expected fragment size: 3609 bp (indicated
by a red arrow). 1-6: clones 1-6 potential E. coli K-12 BW25113 pSW3; C-: negative control (no template was
used for restriction); M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
Figure S11. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 BW25113
pSW3_lacIq clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for E. coli
K-12 BW25113 pSW3_lacIq is 2010 bp while for E. coli K-12 BW25113 pSW3 is 1637 bp. 2-9: clones 1-8 potential
E. coli K-12 BW25113 pSW3_lacIq; 10: E. coli K-12 BW25113 pSW3; Std (1 and 11): molecular weight marker
(Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5 6 7 8 9 M
M 1 2 3 4 5 6 C- M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
187
Figure S12. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 BW25113
pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for E. coli
K-12 BW25113 pSW3_lacI+ is 2010 bp. 2-6: clones 1-5 potential E. coli K-12 BW25113 pSW3_lacI+; Std (1 and 7):
molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S13. Gel electrophoresis of XhoI- and MssI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD clones for verification of plasmid pACG_araBAD presence. Expected fragment
sizes: 2678 and 4559 bp (indicated by red arrows). 1-7: clones 1-7 potential E. coli K-12 NEB5α pACG_araBAD;
M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
Figure S14. Gel electrophoresis of XhoI- and NheI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD_thrA clones for verification of plasmid pACG_araBAD_thrA presence. Expected
fragment sizes: 3658 and 6022 bp (indicated by red arrows). 1-6: clones 1-6 potential E. coli K-12 NEB5α
pACG_araBAD_thrA; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
M 1 2 3 4 5 6 7 M
M 1 2 3 4 5 6 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
188
Appendix
Figure S15. Gel electrophoresis of XhoI- and NheI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD_ilvIH clones for verification of plasmid pACG_araBAD_ilvIH presence. Expected
fragment sizes: 3414 and 6022 bp (indicated by red arrows). 1-6: clones 1-6 potential E. coli K-12 NEB5α
pACG_araBAD_ilvIH; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
Figure S16. Gel electrophoresis of XhoI- and NheI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD_ilvC clones for verification of plasmid pACG_araBAD_ilvC presence. Expected
fragment sizes: 2671 and 6022 bp (indicated by red arrows). 1-5: clones 1-5 potential E. coli K-12 NEB5α
pACG_araBAD_ilvC; C-: negative control (no template); M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb)).
Figure S17. Gel electrophoresis of XhoI- and NheI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD_ilvA clones for verification of plasmid pACG_araBAD_ilvA presence. Expected
fragment sizes: 2740 and 6022 bp (indicated by red arrows). 1-6: clones 1-6 potential E. coli K-12 NEB5α
pACG_araBAD_ilvA; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
1 2 3 4 M M 5 6
M C- 1 2 3 4 5
1 2 3 4 5 6 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
189
Figure S18. Gel electrophoresis of NotI- and MssI-restriction products from plasmid preparations of potential E.
coli K-12 NEB5α pACG_araBAD_leuA clones for verification of plasmid pACG_araBAD_leuA presence. Expected
fragment sizes: 3040 and 5749 bp (indicated by red arrows). 1-3: clones 1-3 potential E. coli K-12 NEB5α
pACG_araBAD_leuA; C: plasmid pACG_araBAD; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder
(0.1-10.0 kb)).
Figure S19. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_araBAD_ilvGM clones by using primers InFusion1_seq_F4 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_araBAD_ilvGM is 2127 bp (indicated by a red arrow). 2-9: clones 1-8
potential E. coli K-12 NEB5α pACG_araBAD_ilvGM; Std (1 and 10): molecular weight marker (Quick-Load® 2-Log
DNA Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
M C 1 2 3 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
190
Appendix
Figure S20. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR product
size for E. coli K-12 NEB5α pACG_XylSPm is 2453 bp (indicated by red rectangles). 1-20: clones 1-20 potential E.
coli K-12 NEB5α pACG_XylSPm; Std (1 and 10): M: molecular weight marker (Quick-Load® 2-Log DNA Ladder
(0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
M 1 2 3 4 5 6 7 8 9 10 M
M 11 12 13 14 15 16 17 18 19 20 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
191
Figure S21. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_ilvIH clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_ilvIH is 4652 bp (indicated by the red arrow). 1-30: clones 1-
30 potential E. coli K-12 NEB5α pACG_ XylSPm_ilvIH; M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
M 1 2 3 4 5 6 7 8 9 10 M
M 11 12 13 14 15 16 17 18 19 20 M
M 21 22 23 24 25 26 27 28 29 30 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
192
Appendix
Figure S22. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_leuA clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_leuA is 4005 bp (indicated by the red arrow). 1-10: clones 1-
10 potential E. coli K-12 NEB5α pACG_ XylSPm_leuA; M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
Figure S23. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_ilvC clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_ilvC is 3909 bp (indicated by the red arrow). 1-10: clones 1-10
potential E. coli K-12 NEB5α pACG_ XylSPm_ilvC; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder
(0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
M 1 2 3 4 5 6 7 8 9 10 M
M 1 2 3 4 5 6 7 8 9 10 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
193
Figure S24. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_ilvA clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_ilvA is 3978 bp (indicated by the red arrow). 1-10: clones 1-
10 potential E. coli K-12 NEB5α pACG_ XylSPm_ilvA; M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
Figure S25. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_ilvBN clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_ilvBN is 4416 bp (indicated by the red arrow). 1-10: clones 1-
10 potential E. coli K-12 NEB5α pACG_XylSPm_ilvBN; M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
M 1 2 3 4 5 6 7 8 9 10 M
M 1 2 3 4 5 6 7 8 9 10 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
194
Appendix
Figure S26. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 NEB5α
pACG_XylSPm_thrA clones by using primers InFusion1_seq_F17 and InFusion1_seq_R4. The expected PCR
product size for E. coli K-12 NEB5α pACG_XylSPm_thrA is 4896 bp (indicated by the red arrow). 1-10: clones 1-
10 potential E. coli K-12 NEB5α pACG_XylSPm_thrA; M: molecular weight marker (Quick-Load® 2-Log DNA
Ladder (0.1-10.0 kb). Molecular weights of the DNA standard are shown in kb.
Figure S27. Gel electrophoresis of EcoRI-restriction products from plasmid preparations of potential E. coli K-12
BW25113 geneX:kanR mutants for verification of plasmid pKD46 curation. Expected fragment size by pKD46
presence: 1509 and 4820 bp (indicated by red arrows). In case of pKD46 absence, no band is expected. 1: E. coli
K-12 BW25113 leuA:kanR; 2: E. coli K-12 BW25113 ilvA:kanR; 3: E. coli K-12 BW25113 ilvB:kanR; 4: E. coli K-12
BW25113 ilvC:kanR; 5: E. coli K-12 BW25113 thrA:kanR; 6: E. coli K-12 BW25113 thrB:kanR; 7: E. coli K-12
BW25113 thrC:kanR; 8: E. coli K-12 BW25113 ilvI:kanR; C-: negative control (E. coli K-12 BW25113); C+: positive
control (pKD46); M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
M 1 2 3 4 5 6 7 8 9 10 M
M 1 2 3 4 5 6 7 C- C+ M 8 C+ M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
195
Figure S28. Gel electrophoresis of EcoRI-restriction products from plasmid preparations of potential E. coli K-12
BW25113 geneX:kanR pCP20 mutants for verification of plasmid pCP20 presence. Expected fragment size by
pCP20 presence: 3454 and 6043 bp (indicated by red arrows). In case of pCP20 absence, no band is expected.
1-2: E. coli K-12 BW25113 thrA:kanR pCP20 clones 1-2; 3-4: E. coli K-12 BW25113 thrB:kanR pCP20 clones 1-2;
5-6: E. coli K-12 BW25113 thrC:kanR pCP20 clones 1-2; 7-8: E. coli K-12 BW25113 leuA:kanR pCP20 clones 1-2;
9-10: E. coli K-12 BW25113 ilvC:kanR pCP20 clones 1-2; 11-12: E. coli K-12 BW25113 ilvA:kanR pCP20 clones 1-
2; 13-14: E. coli K-12 BW25113 ilvB:kanR pCP20 clones 1-2; 15-18: E. coli K-12 BW25113 ilvI:kanR pCP20 clones
1-4; C+: positive control (pCP20); C-: negative control (E. coli K-12 BW25113); M: molecular weight marker
(Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
M 1 2 3 4 5 6 7 8 C- C+ M
M 9 10 11 12 13 14 C- C+ M 15 16 M 17 18
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
196
Appendix
Figure S29. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 BW25113
ΔgeneX pCP20 clones by using primers mentioned in Table 22 to verify removal of the kanamycin resistance
cassette from the genome. The expected PCR product size for each mutant is shown in Table 22 (indicated with
red squares for kanS variants). 1-2: E. coli K-12 BW25113 ΔthrB pCP20 (kanS) clones 1-2; 3: E. coli K-12
BW25113 thrB:kanR pCP20 (kanR); 4-5: E. coli K-12 BW25113 ΔthrC pCP20 (kanS) clones 1-2; 6: E. coli K-12
BW25113 thrC:kanR pCP20 (kanR); 7-8: E. coli K-12 BW25113 ΔilvA pCP20 (kanS) clones 1-2; 9: E. coli K-12
BW25113 ilvA:kanR pCP20 (kanR); 10-11: E. coli K-12 BW25113 ΔilvB pCP20 (kanS) clones 1-2; 12: E. coli K-12
BW25113 ilvB:kanR pCP20 (kanR); 13-14: E. coli K-12 BW25113 ΔthrA pCP20 (kanS) clones 1-2; 15: E. coli K-12
BW25113 thrA:kanR pCP20 (kanR); 16-17: E. coli K-12 BW25113 ΔleuA pCP20 (kanS) clones 1-2; 18: E. coli K-12
BW25113 leuA:kanR pCP20 (kanR); 19-20: E. coli K-12 BW25113 ΔilvC pCP20 (kanS) clones 1-2; 21: E. coli K-12
BW25113 ilvC:kanR pCP20 (kanR); 22-24: E. coli K-12 BW25113 ΔilvI pCP20 (kanS) clones 1-3; 25: E. coli K-12
BW25113 ilvI:kanR pCP20 (kanR); 26: E. coli K-12 BW25113; C-: no DNA; M: molecular weight marker (Quick-
Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5 6 M M 7 8 9 10 11 12 13 14 15 M
16 17 18 M 19 20 21 22 23 24 25 26 C- M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
197
Figure S30. Gel electrophoresis of EcoRI-restriction products from plasmid preparations of potential E. coli K-12
BW25113 pKD46 clones for verification of plasmid pKD46 presence. Expected fragment sizes: 1509 and 4820 bp
(indicated by red arrows). 1-6: clones 1-6 potential E. coli K-12 BW25113 pKD46. M: molecular weight marker
(Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
Figure S31. Gel electrophoresis of the ilvIH knock-out cassette obtained by PCR under different templates and
buffer conditions. Expected fragment sizes (indicated by red arrows): 1105 bp (pKD3) or 1568 bp (pKD4). 1:
template pKD3, buffer HF plus DMSO; 2: template pKD3, buffer GC plus DMSO; 3: negative control; 4: template
pKD4, buffer HF plus DMSO; 5: template pKD4, buffer GC plus DMSO; M: molecular weight marker (Quick-
Load® 2-Log DNA Ladder (0.1-10.0 kb)).
M 1 2 3 4 5 6 M
1 2 3 4 5 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
198
Appendix
Figure S32. Gel electrophoresis of PCR products obtained from genomic DNA of potential E. coli K-12 BW25113
ilvIH:cmR pKD46 clones by using primers ilvIH_F and ilvIH_R. The expected PCR product size for ilvIH mutants is
1658 bp (indicated by a red arrow) while for wild type is 2844 bp. 1-10: clones 1-10 potential E. coli K-12
BW25113 ilvIH:cmR pKD46; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
Figure S33. Gel electrophoresis of PCR products obtained from genomic DNA of potential E. coli K-12 BW25113
ΔilvIH pCP20 clones by using primers ilvIH_F2 and ilvIH_R. The expected PCR product size for E. coli K-12
BW25113 ΔilvIH pCP20 is 834bp (indicated by a red arrow). 1-4: clones 1-4 potential E. coli K-12 BW25113
ΔilvIH pCP20; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb)).
M 1 2 3 4 5 6 7 8 9 10 M
M 1 2 3 4 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
199
Figure S34. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 BW25113
ΔgeneX pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size
for positive clones is 2010 bp. 1-3: E. coli K-12 BW25113 ΔilvA pSW3_lacI+ clones 1-3; 4-6: E. coli K-12 BW25113
ΔilvBN pSW3_lacI+ clones 1-3; 7-9: E. coli K-12 BW25113 ΔilvC pSW3_lacI+ clones 1-3; M: molecular weight
marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S35. Gel electrophoresis of PCR products obtained by colony PCR of potential E. coli K-12 BW25113
ΔgeneX pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size
for positive clones is 2010 bp. 1-5: E. coli K-12 BW25113 ΔleuA pSW3_lacI+ clones 1-5; 6-9: E. coli K-12
BW25113 ΔthrA pSW3_lacI+ clones 1-4; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0
kb).
M 1 2 3 4 5 6 7 8 9 M
M 1 2 3 4 5 6 7 8 9 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
200
Appendix
Figure S36. Gel electrophoresis of PCR products obtained by colony PCR of 10 potential E. coli K-12 BW25113
pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for
positive clones is 2010 bp. 2-11: E. coli K-12 BW25113 pSW3_lacI+ clones 1-10; 1, 12: molecular weight marker
(Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S37. Gel electrophoresis of PCR products obtained by colony PCR of 10 potential E. coli K-12 BW25113
ΔilvIH pSW3_lacI+ clones by using primers pSW3_R3_seq and pSW3_R6_seq. The expected PCR product size for
positive clones is 2010 bp (indicated by a red arrow). 2-11: E. coli K-12 BW25113 ΔilvIH pSW3_lacI+ clones 1-10;
1, 12: molecular weight marker (GeneRuler™ 1 kb Plus DNA Ladder).
1 2 3 4 5 6 7 8 9 10 11 12
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
201
Figure S38. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔleuA pSW3_lacI+ pACG_araBAD_leuA clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 1795 bp. 1-5: E. coli K-12 BW25113 ΔleuA pSW3_lacI+
pACG_araBAD_leuA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S39. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvC pSW3_lacI+ pACG_araBAD_ilvC clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 1699 bp. 1-5: E. coli K-12 BW25113 ΔilvC pSW3_lacI+
pACG_araBAD_ilvC clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5
M 1 2 3 4 5
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
202
Appendix
Figure S40. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔthrA pSW3_lacI+ pACG_araBAD_thrA clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 2686 bp. 1-5: E. coli K-12 BW25113 ΔthrA pSW3_lacI+
pACG_araBAD_thrA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S41. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvA pSW3_lacI+ pACG_araBAD_ilvA clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 1768 bp. 1-5: E. coli K-12 BW25113 ΔilvA pSW3_lacI+
pACG_araBAD_ilvA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5
M 1 2 3 4 5
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
203
Figure S42. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvBN pSW3_lacI+ pACG_araBAD_ilvBN clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 2206 bp. 1-5: E. coli K-12 BW25113 ΔilvBN pSW3_lacI+
pACG_araBAD_ilvBN clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S43. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 2442 bp. 1-5: E. coli K-12 BW25113 ΔilvIH pSW3_lacI+
pACG_araBAD_ilvIH clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5
M 1 2 3 4 5
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
204
Appendix
Figure S44. Gel electrophoresis of PCR products obtained by colony PCR of 8 potential E. coli K-12 BW25113
ΔilvGM pSW3_lacI+ pACG_araBAD_ilvGM clones by using primers InFusion1_seq_F4 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 2127 bp. 1-8: E. coli K-12 BW25113 ΔilvGM pSW3_lacI+
pACG_araBAD_ilvGM clones 1-8; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S45. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvC pSW3_lacI+ pACG_XylSPm_ilvC clones by using primers InFusion1_seq_F17 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 3909 bp. 1-5: E. coli K-12 BW25113 ΔilvC pSW3_lacI+
pACG_XylSPm_ilvC clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
1 2 3 4 5 6 7 8 M
M 1 2 3 4 5
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
205
Figure S46. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔleuA pSW3_lacI+ pACG_XylSPm_leuA clones by using primers InFusion1_seq_F17 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 4005 bp. 1-5: E. coli K-12 BW25113 ΔleuA pSW3_lacI+
pACG_XylSPm_leuA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S47. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔthrA pSW3_lacI+ pACG_XylSPm_thrA clones by using primers InFusion1_seq_F17 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 4896 bp. 1-5: E. coli K-12 BW25113 ΔthrA pSW3_lacI+ pACG_
XylSPm_thrA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5
1 2 3 4 5 M
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
206
Appendix
Figure S48. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvA pSW3_lacI+ pACG_XylSPm_ilvA clones by using primers InFusion1_seq_F17 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 3978 bp. 1-5: E. coli K-12 BW25113 ΔilvA pSW3_lacI+ pACG_
XylSPm_ilvA clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
Figure S49. Gel electrophoresis of PCR products obtained by colony PCR of 5 potential E. coli K-12 BW25113
ΔilvBN pSW3_lacI+ pACG_XylSPm_ilvBN clones by using primers InFusion1_seq_F17 an InFusion1_seq_R4. The
expected PCR product size for positive clones is 4416 bp. 1-5: E. coli K-12 BW25113 ΔilvBN pSW3_lacI+
pACG_XylSPm_ilvBN clones 1-5; M: molecular weight marker (Quick-Load® 2-Log DNA Ladder (0.1-10.0 kb).
M 1 2 3 4 5
M 1 2 3 4 5
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
207
Figure S50. Sequence determination of plasmid pSW3 by overlapping of sequencing reads. Generated with
Snapgene®.
Figure S51. Sequencing reads against sequence of plasmid pSW3_lacIq for clones 3 and 5 (from top to bottom)
E. coli NEB5α pSW3_lacIq. Generated with Snapgene®.
Figure S52. Sequencing reads against sequence of plasmid pSW3_lacI+ for clone 1 E. coli NEB5α pSW3_lacI+.
Generated with Snapgene®.
Figure S53. Overlapping sequencing reads against sequence of plasmid pACG_araBAD for clone 1 E. coli NEB5α
pACG_araBAD. Generated with Snapgene®.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
208
Appendix
Figure S54. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_leuA for clone 2 E. coli
NEB5α pACG_araBAD_leuA. Generated with Snapgene®.
Figure S55. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_ilvC for clone 4 E. coli
NEB5α pACG_araBAD_ilvC. Generated with Snapgene®.
Figure S56. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_ilvA for clone 5 E. coli
NEB5α pACG_araBAD_ilvA. Generated with Snapgene®.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
209
Figure S57. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_thrA for clone 3 E. coli
NEB5α pACG_araBAD_thrA. Generated with Snapgene®.
Figure S58. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_ilvIH for clone 1 E. coli
NEB5α pACG_araBAD_ilvIH. Generated with Snapgene®.
Figure S59. Overlapping sequencing reads against sequence of plasmid pACG_araBAD_ilvGM for clone 1 E. coli
NEB5α pACG_araBAD_ilvGM. Generated with Snapgene®.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
210
Appendix
Figure S60. Sequencing reads against sequence of plasmid pACG_XylSPm for clone 16 E. coli NEB5α
pACG_XylSPm. Generated with Snapgene®.
Figure S61. Sequencing reads against sequence of plasmid pACG_XylSPm_leuA for clone 4 E. coli NEB5α
pACG_XylSPm_leuA. Generated with Snapgene®.
Figure S62. Sequencing reads against sequence of plasmid pACG_XylSPm_ilvC for clone 1, 3, 6 and 9 (from top
to bottom) E. coli NEB5α pACG_XylSPm_ilvC. Generated with Snapgene®.
Figure S63. Sequencing reads against sequence of plasmid pACG_XylSPm_ilvIH for clone 21 E. coli NEB5α
pACG_XylSPm_ilvIH. Generated with Snapgene®.
Figure S64. Sequencing reads against sequence of plasmid pACG_XylSPm_thrA for clones 2 and 8 (from top to
bottom) E. coli NEB5α pACG_XylSPm_thrA. Generated with Snapgene®.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
211
Figure S65. Sequencing reads against sequence of plasmid pACG_XylSPm_ilvBN for clone 3 E. coli NEB5α
pACG_XylSPm_ilvBN. Generated with Snapgene®.
Figure S66. Sequencing reads against sequence of plasmid pACG_XylSPm_ilvA for clones 4 and 6 (from top to
bottom) E. coli NEB5α pACG_XylSPm_ilvA. Generated with Snapgene®.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
212
Appendix
Figure S67. Verification of plasmid pCP20 curation in strains E. coli K-12 BW25113 ΔleuA, ΔilvC and ΔilvI (A, B)
and E. coli K-12 BW25113 ΔilvA, ΔilvB, ΔthrA, ΔthrB and ΔthrC (C). A single 42 °C treatment was used in (A) and
(C) while a second 42 °C treatment was additionally carried out in (B). For each clone, cell material was in
parallel streaked out in 3 different plates: LB containing 100 µg/mL ampicillin, LB containing 30 µg/mL
kanamycin and LB without antibiotics. Kanamycin sensitivity indicates that kanamycin resistance marker was
successfully removed from the genome. Ampicilin sensitivity indicates loss of plasmid pCP20. Accordingly,
clones growing in LB but not in antibiotic-containing plates have the correct phenotype after complete curation
of pCP20 (indicated in red). Clones growing in LB but partially in LB-ampicillin (indicated in green) did not lose
all copies of pCP20 and need to be temperature-treated again. As a control for selection following strains were
used: E. coli K-12 BW25113 ilvI:kanR (indicated in yellow, kanamycin resistant) and E. coli K-12 BW25113 ΔilvI
pCP20 (indicated in orange, ampicillin resistant).
A B
C C
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
213
Figure S68. Verification of plasmid pKD46 curation after temperature treatment in subclones originating from
clone 1 E. coli K-12 BW25113 ilvIH:cmR pKD46. For each subclone, cell material was in parallel streaked out in 3
different plates: LB containing 100 µg/mL ampicillin (lower-left), LB containing 12.5 µg/mL chloramphenicol
(lower-right) and LB without antibiotics (upper). Ampicilin sensitivity indicates loss of plasmid pKD46.
Accordingly, clones growing in LB and LB-chloramphenicol but not in LB-ampicillin have the correct phenotype
after complete curation of pKD46.
Figure S69. Verification of plasmid pCP20 curation in strain E. coli K-12 BW25113 ΔilvIH. For each clone, cell
material was in parallel streaked out in 3 different plates: LB containing 100 µg/mL ampicillin, LB containing
12.5 µg/mL chloramphenicol and LB without antibiotics. Chloramphenicol sensitivity indicates that
chloramphenicol resistance marker was successfully removed from the genome. Ampicilin sensitivity indicates
loss of plasmid pCP20. Accordingly, clones growing in LB but not in antibiotic-containing plates have the correct
phenotype after complete curation of pCP20.
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
214
Appendix
Figure S70. Calculated molar ratios of the ncBCAA with respect to the canonical counterparts in the intracellular
soluble protein fraction (A) and inclusion body fraction (B) from samples taken 3.5 h after IPTG induction of E.
coli K-12 BW25113 pSW3_lacI+ cultivations in shake flasks under cultivation conditions subjected to pulses of
different pyruvate concentration. When cultivation achieved OD600nm of 0.3, pyruvate was supplemented in a
pulse-based manner each 30 minutes for a time period of 2.5h, this is a total of 6 pyruvate pulses. Depending
on the cultivation pyruvate pulses were added up to 0, 100, 250, 500, 1000, 2000 and 5000 mg/L. Induction
with 0.5 mM IPTG was performed when cultures reached OD600nm of 0.6.
Figure S71. Molar concentrations of norvaline (blue bars) and norleucine (red bars) normalized to PPI mass
present in the inclusion body fraction at 3h after induction of E. coli K-12 BW25113 pSW3_lacI+ cultivation in a
10 mL mini-reactor under standard cultivation conditions (-) or under conditions triggering ncBCAA production,
i.e. pyruvate pulses combined with O2 limitation (+).
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
215
Figure S72. Calculated molar ratios of the ncBCAA with respect to the canonical counterparts in the intracellular
soluble protein fraction (A) and inclusion body fraction (B) at 3h after induction of E. coli K-12 BW25113
pSW3_lacI+ cultivation in a 10mL mini-reactor under standard cultivation conditions (-) or under conditions
triggering ncBCAA production, i.e. pyruvate pulses combined with O2 limitation (+).
Figure S73. OD600nm (A) and CDW (B) over time after induction of E. coli K-12 BW25113 pSW3_lacI+ cultivation
(WT E. coli) in a 15L reactor under standard cultivation conditions (STD) or under conditions triggering ncBCAA
accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-O2).
A B
A B
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
216
Appendix
Table S8. Estimated recombinant mini-proinsulin concentrations (µg PPI/OD0.25) for each strain under the
different tested cultivation conditions in the 10mL mini-reactor. This data correspond to the mini-reactor plate
shown in Table 26. Two values are shown for each strain and condition: the first refers to standard cultivation
conditions while the second applies to cultivation conditions triggering ncBCAA formation, i.e. pyruvate pulsing
and O2 limitation. Depending on the sample, analysis was done at different time points after induction: 2h AI
(2h after induction) or 3.5h AI (3.5h after induction).
Strain
L-arabinose concentration (%)
0
0.1
0.2
0.4
0.8
1.6
E. coli K-12 BW25113
pSW3_lacI+
0.92/0.91
(3.5h AI)
0.82/0.78
(2h AI)
-
-
-
-
-
leuA-tunable E. coli
-
0.95/0.91
(3.5h AI)
0.97/0.99
(3.5h AI)
1.02/1.07
(3.5h AI)
-
-
ilvC-tunable E. coli
-
-
-
0.80/0.73
(2h AI)
0.80/0.82
(2h AI)
0.75/0.79
(2h AI)
thrA-tunable E. coli
-
-
-
0.76/0.84
(3.5h AI)
0.70/0.80
(3.5h AI)
0.80/0.76
(3.5h AI)
Table S9. Estimated recombinant mini-proinsulin concentrations (µg PPI/OD0.25) 2h after induction for each
strain under the different tested cultivation conditions in the 10mL mini-reactor. This data correspond to the
mini-reactor plate shown in Table 27. Two values are shown for each strain and condition: the first refers to
standard cultivation conditions while the second applies to cultivation conditions triggering ncBCAA formation,
i.e. pyruvate pulsing and O2 limitation.
Strain
L-arabinose concentration (%)
0
0.05
0.2
0.8
E. coli K-12 BW25113
pSW3_lacI+
0.79/0.79
-
-
-
ilvIH-tunable E. coli
-
0.82/0.64
-
0.79/0.82
ilvA-tunable E. coli
-
0.91/0.89
0.95/0.97
0.78/0.92
ilvBN-tunable E. coli
-
0.66/0.76
0.78/0.68
0.68/0.60
ilvGM-tunable E. coli
-
0.66/-
0.65/0.71
0.59/0.77
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
217
Figure S74. Calculated molar ratios of the ncBCAA with respect to the canonical counterparts in the intracellular
soluble protein fraction from samples taken 2-3.5 h after IPTG induction of cultivations of ilvC-tunable E. coli
(A), leuA-tunable E. coli (B), thrA-tunable E. coli (C), ilvIH-tunable E. coli (D), ilvA-tunable E. coli (E), ilvBN-
tunable E. coli (F) and ilvGM-tunable E. coli (G) in a 10mL mini-reactor subjected to different L-arabinose
concentrations and cultivation modes. Two cultivation modes were tested: standard cultivation conditions (-)
and conditions triggering ncBCAA production, i.e. pyruvate pulses combined with O2 limitation (+). Strain E. coli
A B
C D
E F
G
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
218
Appendix
K-12 BW25113 pSW3_lacI+ (indicated as “wild type E. coli” in the chart) was employed as a control for
comparison.
Figure S75. Calculated molar ratios of the ncBCAA with respect to the canonical counterparts in the inclusion
body fraction from samples taken 2-3.5 h after IPTG induction from cultivations of ilvC-tunable E. coli (A), leuA-
tunable E. coli (B), thrA-tunable E. coli (C), ilvIH-tunable E. coli (D), ilvA-tunable E. coli (E), ilvBN-tunable E. coli
(F) and ilvGM-tunable E. coli (G) in a 10mL mini-reactor subjected to different L-arabinose concentrations and
cultivation modes. Two cultivation modes were tested: standard cultivation conditions (-) and conditions
A B
C D
E F
G
F
Molecular genetic approaches to decrease mis-incorporation of non-canonical
Ángel Córcoles García branched chain amino acids into a recombinant protein in Escherichia coli
Appendix
219
triggering ncBCAA production, i.e. pyruvate pulses combined with O2 limitation (+).Strain E. coli K-12 BW25113
pSW3_lacI+ (indicated as wild type E. coli in the chart) was employed as a control for comparison.
Figure S76. OD600nm (A) and CDW (B) over time after induction of different E. coli cultivations in a 15L reactor
under cultivation conditions triggering ncBCAA accumulation, i.e. pyruvate pulsing and oxygen limitation (PYR-
O2). Indicated in the legend, “WT E.coli” refers to the wild type strain E. coli K-12 BW25113 pSW3_lacI+,
“ilvGM-tunable E. coli” alludes to strain E. coli K-12 BW25113 pSW3_lacI+ pACG_araBAD_ilvGM and “ilvIH-
tunable E. coli” corresponds with strain E. coli K-12 BW25113 ΔilvIH pSW3_lacI+ pACG_araBAD_ilvIH.
Figure S77. Molar concentrations of norvaline (A, C) and norleucine (B, D) normalized to PPI mass present in the
inclusion body fraction calculated over time after induction of different E. coli cultivations in a 15L reactor
under conditions triggering ncBCAA accumulation; i.e. pyruvate pulsing and oxygen limitation (PYR-O2), in a big
(A, B) and reduced y-axis scale (C, D). Indicated in the legend, “WT E.coli” refers to the wild type strain E. coli K-
12 BW25113 pSW3_lacI+, “ilvGM-tunable E. coli” alludes to strain E. coli K-12 BW25113 pSW3_lacI+
pACG_araBAD_ilvGM and “ilvIH-tunable E. coli” corresponds with strain E. coli K-12 BW25113 ΔilvIH
pSW3_lacI+ pACG_araBAD_ilvIH. Orange arrows indicate time points where 1 g/L pyruvate pulse combined with
5 min O2 limitation was applied.
A B
A B
C D
Molecular genetic approaches to decrease mis-incorporation of non-canonical
branched chain amino acids into a recombinant protein in Escherichia coli Ángel Córcoles García
220
Appendix
ggcatgcatttacgttgacaccatcgaatggcgcaaaacctttcgcggtatggcatgatagcgcccggaagagagtcaattcagggtggtgaatg
tgaaaccagtaacgttatacgatgtcgcagagtatgccggtgtctcttatcagaccgtttcccgcgtggtgaaccaggccagccacgtttctgcga
aaacgcgggaaaaagtggaagcggcgatggcggagctgaattacattcccaaccgcgtggcacaacaactggcgggcaaacagtcgttgctga
ttggcgttgccacctccagtctggccctgcacgcgccgtcgcaaattgtcgcggcgattaaatctcgcgccgatcaactgggtgccagcgtggtgg
tgtcgatggtagaacgaagcggcgtcgaagcctgtaaagcggcggtgcacaatcttctcgcgcaacgcgtcagtgggctgatcattaactatccg
ctggatgaccaggatgccattgctgtggaagctgcctgcactaatgttccggcgttatttcttgatgtctctgaccagacacccatcaacagtattat
tttctcccatgaagacggtacgcgactgggcgtggagcatctggtcgcattgggtcaccagcaaatcgcgctgttagcgggcccattaagttctgt
ctcggcgcgtctgcgtctggctggctggcataaatatctcactcgcaatcaaattcagccgatagcggaacgggaaggcgactggagtgccatgt
ccggttttcaacaaaccatgcaaatgctgaatgagggcatcgttcccactgcgatgctggttgccaacgatcagatggcgctgggcgcaatgcgc
gccattaccgagtccgggctgcgcgttggtgcggatatctcggtagtgggatacgacgataccgaagacagctcatgttatatcccgccgttaacc
accatcaaacaggattttcgcctgctggggcaaaccagcgtggaccgcttgctgcaactctctcagggccaggcggtgaagggcaatcagctgtt
gcccgtctcactggtgaaaagaaaaaccaccctggtgcccaatacgcaaaccgcctctccccgcgcgttggccgattcattaatgcagctggcac
gacaggtttcccgactggaaagcgggcagtgagcgcaacgcgcgattgaggaccgcatctgcgctgggcttggcttcgccgagcatcagcgcgt
cagtgccgtgcatcacgacaccgacaacctgcacatccatatcgccatcaacaagattcacccgacccgaaacaccatccatgagccgtatcgg
gcctaccgcgccctcgctgacctctgcgcgacgctcgaacgggactacgggcttgagcgtgacaatcacgaaacgcggcagcgc
Figure S78. Nucleotide sequence of the genetic region of E. coli K-12 BW25113 genome containing the gene lacI
and its corresponding lacI+ promoter variant. Blue: promoter sequence corresponding to the lacI promoter. Red:
single base offering lacI+ phenotype. Green: coding region of gene lacI.
s
Recent dissertations at Bioprocess Engineering, TU Berlin
1. Sarah Kamel Hassan (2019) Enzymatic synthesis of α-D-pentofuranose-1-phosphate as a universal
glycosyl donor for modified nucleosides synthesis using thermostable nucleoside phosphorylase
2. Funda C. Erterm (2019) Life cycle assessment and modelling approaches as a combined evaluation
tool for sustainable control strategies at biogas plants
3. Emmanuel Anane (2019) Model-based strategies for scale-down studies in fed-batch cultivation of
Escherichia coli expressing proinsulin
4. Martin Kögler (2018) Advanced Raman spectroscopy for bioprocess monitoring
5. Anna Maria Marbá (2018) Monitoring of the single-cell morphology for the evaluation of microbial
eukaryotic bioprocesses
6. Juan Antonio Arzate (2018) Modelling and simulation of biogas production based on anaerobic
digestion of energy crops and manure
7. Anika Bockisch (2018) Mobile multi-parameter measurements for the dynamic analysis of
gradients in brewing vessels
8. Elvis Legala (2018) Reconstruction of the lanthipeptide ruminococcin-A biosynthesis machinery in
Escherichia coli and structural characterization
9. Ekaterina Osmekhina (2018) Quantitative approach for detection of RNA and proteins by sandwich
hybridization technology
10. Anja Lemoine (2017) Impact of oscillations in substrate and oxygen availability on
Corynebacterium glutamicum cultivations
11. Basant El Kady (2017) Application of the enzyme controlled glucose feeding for the production of
recombinant proteins. Starter cultures and autoinduction
12. Julia Glazyrina (2017) Endophytes as source of polypeptide and polyketide antibiotics fengycin and
bacillaene. Process development for fengycin production
13. Florian Glauche (2017) Platform technologies for automated bioprocess development
14. Eva Brand (2017) Growth rate dependent impact of oscillating conditions on Escherichia coli
physiology
